Assessment of Service Quality and Trust of E-Public Transportation in Doha Qatar

Abstract

This study examined the relationship between service quality, trust, and passenger satisfaction in sustainable e-public transportation, using the Doha Metro in Qatar as a case study. Despite its advanced automation, the metro faces low adoption, with less than 30% of the economically active population utilizing it. To address this, this study integrated the Social Exchange Theory (SET) and the SERVQUAL RATER model with machine learning techniques to assess commuter perceptions and satisfaction. Neural network and Long Short-Term Memory (LSTM) models outperformed traditional statistical methods, offering enhanced predictive accuracy. Based on the 319 survey responses, key service quality factors were identified, emphasizing customer experience (NI = 100.0%), security (NI = 99.9%), and service reliability (NI = 90.8%). Findings suggested that improving affordability and dynamic pricing could increase metro ridership while reducing private vehicle reliance. Additionally, predictive maintenance and crisis management strategies are recommended to enhance service reliability. This study contributes to sustainable urban mobility by providing data-driven recommendations for efficient and environmentally friendly e-public transportation. Policymakers and urban planners can utilize these insights to improve commuter satisfaction and transit system adoption. Future research may explore multi-city comparisons and hybrid modeling techniques for further refinement.

1. Introduction

Public transportation networks are vital for fostering sustainable urban development by providing effective mobility solutions while minimizing environmental impacts [1]. Batarce et al. [2] highlighted the significance of addressing urbanization issues such as transportation bottlenecks, air quality degradation, and resource management in creating future cities. The Doha Metro Rail System, which opened in 2019, marked a pivotal moment in Qatar’s commitment to sustainable infrastructure, enhancing urban connections, and advancing the country’s economic and environmental objectives [3]. As a fundamental component of Qatar’s National Vision 2030, the metro system sought to reduce car dependency, promote public transportation use, and contribute to a greener, more efficient urban mobility framework. With a population of approximately 3.08 million, 82% of whom reside in Doha [4], the Doha Metro transportation system has become a crucial element of daily commuting. Since its launch, the system has recorded over 200 million riders, averaging more than 100,000 passengers daily, solidifying its role as the city’s transportation backbone [3,5].

Understanding commuter satisfaction is necessary for the ongoing success and enhancement of systems for automated public transportation like the Doha Metro. This electric public transportation system utilizes a state-of-the-art GoA4 system, fully automated and driverless. Promoting sustainable e-public transportation, the optimum operation does not require any staff to monitor the whole system. This overall e-system aims to reduce the carbon footprint and provide better service all while using regenerative braking systems. However, despite its launch in 2022, there is still a need for further assessment because there are less than 30% of economically active population travelers utilizing this e-public transportation [6]. In light of this, the Doha Metro in Qatar requires particular attention due to its unique socio-economic circumstances. Despite Qatar’s enormous financial resources and modern infrastructure, the country’s elevated vehicle ownership and large foreign workforce make it difficult to promote metro use. Improving commuter satisfaction through service integration, cultural barrier removal, and first- and last-mile connectivity is essential to achieving sustainable urban e-mobility.

Passenger satisfaction relies on service quality, which could be assessed via the five popular domains—reliability, tangibles, accessibility, empathy, and assurance [7], as well as other factors such as pricing, trust, social exchange, and overall service excellence [8,9]. Research has shown that superior service quality fosters positive user perceptions, boosts ridership, and ensures the system’s sustainability. For instance, studies conducted during the COVID-19 pandemic highlighted the significance of service quality in maintaining customer satisfaction during health-related challenging times, particularly in public utility vehicle services [8]. Likewise, behavioral research grounded in the Social Exchange Theory and the theory of planned behavior has illuminated passenger perceptions and satisfaction with public transportation throughout the pandemic [9]. Despite the importance of these factors, there has been limited research assessing passenger satisfaction in Middle Eastern e-public transit systems using advanced analytical techniques.

This study sought to fill current gaps in urban e-rail transit research by utilizing a machine learning ensemble to forecast passenger satisfaction through service quality assessment, specifically in the Doha Metro. Grounded in the Social Exchange Theory (SET), which posits that satisfaction is influenced by a balance of perceived benefits and costs [10], this study integrated key aspects of service quality through the RATER SERVQUAL [7] to identify the critical factors affecting customer perceptions. To achieve this, this study utilized advanced machine learning methodologies, particularly neural network and Long Short-Term Memory (LSTM) models, to analyze and evaluate passenger satisfaction and service quality.

Neural networks, which are known for their ability to model complex nonlinear relationships, effectively identify intricate patterns between service quality dimensions and passenger satisfaction metrics [11]. Meanwhile, LSTM networks, a specialized type of recurrent neural network (RNN), are well suited for analyzing sequential data, making them an ideal choice for this study to forecast overall satisfaction [12]. LSTM networks have demonstrated strong capabilities in capturing temporal patterns, such as variations in ridership trends and the evolution of customer perceptions over time. This aligns with recent research by Wang et al. [13], who explored the application of machine learning models for short-term passenger flow prediction in urban rail transit. Their study emphasized the importance of accurate short-term demand forecasting in implementing dynamic control strategies and minimizing service disruptions. By integrating LSTM-based modeling, this study built upon prior research to provide a more comprehensive understanding of ridership trends and service quality within the Doha Metro system.

Compared to the existing works of literature dominated by multivariate analyses like structural equation modeling (SEM), machine learning algorithms have been considered in the recent generation of studies due to their accuracy and ability to delineate the limitations of SEM. As explained in the study of Fan et al. [14] and Cortez et al. [15], SEM would have complications when dealing with large, integrated models—which may present internal misspecifications that need a shadow analysis using a Phantom Variable to underscore significant findings. Moreover, a farther exogenous latent that aims to assess a target variable indirectly may have lower significance with a larger model from integrated frameworks. Woody [16] further confirms this, implying that these larger models may need supporting analysis to decipher more distinct significant measurements to identify accurate findings. To this, studies like Jamshidi et al. [17] proved how different machine learning techniques could provide similar outputs or better analytical results with higher accuracy by comparison. Ong et al. [18] further confirmed this, implicating that there is better output when using random forest and neural network algorithms, but other machine learning algorithms could be implemented.

The implementation of these techniques, neural networks and LSTM, was carried out using the MATLAB R2024b software, a powerful platform for numerical computation and machine learning applications [12]. MATLAB’s extensive libraries, including its neural network and deep learning toolbox, streamline the design, training, and evaluation of models. These capabilities ensure precise and reproducible analysis [15], which is critical for uncovering actionable insights and informing data-driven decisions to enhance service quality and passenger satisfaction.

These findings contribute to the increasing body of knowledge on urban e-mobility by providing a comprehensive and creative approach to evaluating public transportation systems in fast-emerging countries such as Qatar. It emphasizes the possibility of incorporating new technologies into transportation planning and operations by making concrete recommendations for improving service quality for sustainable e-public transportation that other countries may adopt. Furthermore, it emphasizes the significance of adjusting public transportation systems to changing user expectations [12], ultimately promoting sustainable urban development and the broader aims of Qatar’s National Vision 2030.

This study analyzes commuter satisfaction and service quality in the Doha Metro, which aligns with SDG 9 (Industry, Innovation, and Infrastructure), SDG 11 (Sustainable Cities and Communities), SDG 13 (Climate Action), and SDG 8 (Decent Work and Economic Growth). It promotes efficient transport infrastructure and innovation (SDG 9), improves public transportation accessibility and urban livability (SDG 11), and supports environmental sustainability by lowering carbon emissions (SDG 13). Furthermore, enhancing connectivity and commuting experience promotes economic growth and labor productivity (SDG 8). The findings offer data-driven insights for lawmakers and urban planners, resulting in a more efficient, environmentally friendly, and equitable public transportation system that supports Qatar’s National Vision 2030.

The paper is arranged as follows: (1) Introduction, covering the importance of public transport, an overview of the Doha Metro, challenges and need for study, passenger satisfaction and service quality, research gaps, study contributions, and alignment with Sustainable Development Goals (SDGs). Second, (2) related studies and conceptual framework, followed by (3) study methodology including data collection and participants, questionnaire and data preprocessing, and the optimization of algorithms. The (4) results, including validity, comparisons between algorithms, and the generation of variable importance were aligned, and (5) the discussion, including theoretical contribution and practical and managerial implications are presented. Lastly, (6) the conclusion, including limitations and future research are presented.

2. Conceptual Framework

2.1. Theoretical Background

Promoting the use of public transit, reducing traffic, facilitating sustainable urban mobility and improving air quality all depend on metro rail systems. Systems like the Delhi Metro’s performance demonstrate how reliable, clean, and easily available services have raised customer satisfaction in developing countries [19]. The same is true for São Paulo Metro’s systems [20]. According to Aghajanzadeh et al. [21], the service quality of the metro is crucial when there is conflict such as the COVID-19 pandemic. In order to preserve ridership and satisfaction, which are essential for long-term sustainability, their analysis focused on qualities including cleanliness, health safety, and dependability [21]. Similarly, Ismael et al. [22] examined how the perceptions of public transport service quality during the COVID-19 pandemic varied across different user groups, emphasizing the importance of addressing passenger expectations to maintain ridership and ensure long-term viability.

However, cultural inclinations toward private vehicles provide a challenge to the Riyadh Metro, which represents a significant investment in public transit in the Middle East. It was suggested that raising ridership and advancing sustainability depend heavily on improving customer satisfaction through particular tactics [23]. Furthermore, De Oña et al. [24] highlighted that private vehicle users often exhibit different attitudes toward public transport service quality, satisfaction, and overall willingness to switch, reinforcing the need for targeted strategies to encourage public transit adoption.

Customer satisfaction is explained by a variety of theories, such as the Expectancy–Disconfirmation Paradigm, Value Perception Theory, Equity Theory, and Attribute Theory [25]. SET [19,26] and RATER SERVQUAL [10,13] were used for this study, highlighting how important service quality is to client happiness. It could be posited that SET and RATER SERVQUAL were most likely chosen by related studies for their multidimensional approach, direct emphasis on service quality factors, and capacity to capture a more nuanced view of customer satisfaction, which other frameworks lack. The other theories, while useful in certain settings, have limits in addressing the complexities of service satisfaction or focusing primarily on specific components of the commuter experience.

SET claims that interpersonal interactions, which are impacted by social and economic consequences, determine satisfaction levels, which makes it a perfect framework for evaluation [26]. According to the application of SET, adopting SET is required because of an invisible and ambiguous kind of reciprocal accountability. One person will feel compelled to return the kindness later on if they receive something in return; this transaction can take many various forms [10]. In e-public transportation, various forms of exchange influence commuter satisfaction and system sustainability. Passengers expect efficiency, reliability, and comfort in return for their fare, with service quality directly impacting trust and long-term ridership [7]. It could be implicated that a well-maintained system fosters loyalty, while poor service leads to dissatisfaction [9]. Governments invest in metro infrastructure, anticipating public cooperation and responsible usage, yet low adoption may result in reduced funding [8]. The automated Doha Metro enhances efficiency, but passengers must adapt to digital systems, where ease of use boosts satisfaction and complexity discourages adoption [1].

Lastly, first- and last-mile connectivity plays a crucial role in public transport usage—seamless integration improves accessibility, while weak connections push commuters toward private vehicles [13]. As a result, both internal assessments and interpersonal interactions influence how service quality is perceived in metro systems, underscoring its significance in attaining satisfaction for users and advancing sustainable urban e-mobility [8]. The RATER SERVQUAL model is highly advantageous in evaluating and improving service quality in e-public transportation. Its alignment with SET, its ability to address key metro system challenges, and its potential for integration with advanced analytics make it an effective tool for ensuring commuter satisfaction and fostering sustainable urban mobility [7,13]. By systematically assessing service quality dimensions, transit authorities can enhance passenger trust, increase ridership, and promote a more sustainable transportation system in cities like Doha.

2.2. Social Exchange Theory

Particularly when taking reciprocal costs and benefits into account, SET provides an essential basis for understanding decision-making and interaction dynamics. The concept, which has its roots in psychology and sociology, posits that people use a cost–benefit analysis to assess the worth of transactions or relationships in order to maximize the benefits while lowering the costs [10,26]. This conceptual framework is commonly used in transportation research [9] to examine how people respond to new services and systems, such as customer satisfaction [27].

SET is utilized in Qatar’s autonomous train network to assess community satisfaction as well as participation in eco-friendly mobility programs. Autonomous train network construction is a critical constituent of Qatar’s eco-friendly development plan, which is consistent with the country’s 2030 development framework, which emphasizes the advancement of sustainable urban transportation, the alleviation of traffic congestion, and the enhancement of environmental preservation [28]. However, its success depends on how well liked and valuable its consumers think it is. Researchers can look at how the community weighs benefits like cost savings, convenience, and environmental contributions against potential costs like savings in costs, ease of use, and ecological contributions. SET was created and has been shown to be an effective paradigm for analysis in this regard [9,29].

According to Cahigas et al.’s study [9], SET might be one of the frameworks used to gauge commuter satisfaction with public transit. This is due to their implication that, in order to maintain consumer intention and the ongoing use of public transportation services, the management and exploitation of resources should be reciprocated in a social exchange, highlighting the significance of reciprocal benefits between commuters and transport operators in guaranteeing commuter satisfaction. This study used SET to analyze the factors influencing Qatar residents’ satisfaction with Qatar’s autonomous train network and assess the expatriate communities’ perceptions of its value, equity, and mutuality, influencing their willingness to utilize it. This study elucidates the broader consequences of customer satisfaction in order to establish viable urban transportation strategies in Qatar.

Making a site or location more accessible is the definition of accessibility in the context of urban geography and economics [30]. This is among the transportation system’s most significant results. It quantifies the geographic advantage that a zone or a place has over different areas or regions, claim Biosca and Stepniak [31]. Public transportation that operates effectively improves access to other services [32]. Furthermore, this study considers passengers’ opinions of convenience, security, and the Doha Metro’s safety to serve as safety markers. The environmental impact of public transportation and the choice of travel mode are directly impacted by safety, as research has conclusively shown. According to Friman et al. [33], safety perceptions are most strongly correlated with quality categories that offer high levels of comfort and informative value. Views of travel safety in general and the impression of accessibility through perceptions of safety rely on these connections. Comfortable and knowledgeable commuters believe they are easier to approach because they feel safer [9]. A review of the literature claims that one of the aspects of service that is most closely linked to customer satisfaction and public transportation loyalty is safety [34]. In-vehicle crowding significantly affects commuters’ perceptions of public transit, as the presence of overcrowded vehicles is often viewed as a probable intrusion to safety and personal space. This overcrowding is closely associated with overall comfort and a sense of security during travel. Therefore, the hypothesis proposed was as follows:

H1. 

The accessibility measures of the Doha Metro significantly impact passengers’ trust, leading to increased satisfaction among commuters.

H2. 

The safety protocols of the Doha Metro significantly enhance passengers’ trust, leading to increased satisfaction among commuters.

Atombo and Dzigbordi Wemegah [35] came to the conclusion that affordability, or economic benefit, was one factor influencing reliance on public transportation. Inconsistent service, complicated navigation, and insufficient safety and security features negatively affect passenger satisfaction, reduce the likelihood of use, and threaten the financial viability of the system. Cahigas et al. [9] observed that economic benefits play a significant role in building trust in public transportation, with passengers perceiving affordable fares and the easy management of transportation allowances as crucial factors. Their study highlighted how those reasonable fares, especially during difficult times like the COVID-19 pandemic, fostered greater confidence in the system. Enhancing public transportation quality is one of the most important ways to reduce inequality in public transportation, according to Wang et al. [36]. Initiatives like reduced tickets and fare passes for public transportation serve to reduce social disadvantages by making it easier for underprivileged people to obtain essential services and goods. According to Al-Thawadi et al.’s research, which took into account the environmental advantages [37], the majority of the effect categories are more sustainable with the Doha Metro system, provided that 190,000 cars are taken off the road. The role of metro transportation in mitigating the primary environmental impacts of the transportation sector—such as resource depletion, soil degradation, public health risks, air pollution, and global warming—has also been discussed. Therefore, the hypothesis proposed was as follows:

H3. 

Passengers’ trust is significantly impacted by the perceived economic benefit.

Throughout its existence, the Doha Metro has encountered typical difficulties. Since the World Health Organization declared COVID-19 a pandemic on 11 March 2020 [38], Qatar has implemented a clear strategy to mitigate its impact and ensure the best outcomes [39]. Service providers and visitors should align with the primary goal of crisis management, which is to evaluate the effectiveness of government initiatives during the pandemic [40]. Crisis management had a considerable and positive impact on travelers’ impressions of travel in the Philippines, according to Cahigas et al. [9], who examined how people behaved on public transit during the COVID-19 pandemic. Passengers in this study thought that drivers, operators, and transportation service providers should adhere to and carry out the necessary COVID-19 preventive measures. People who adhered to these procedures were able to effectively confine the spread of the COVID-19 disease while they were using public transit. Crowd management presented the Doha Metro with yet another obstacle throughout the fifth day of the 2022 FIFA World Cup, moving 827,000 passengers every day [41], surpassing the average daily ridership of 600,000 during the 2022 FIFA World Cup [42]. Consequently, the Doha Metro gained the trust of 17.4 million commuters who used the Doha Metro to reach the football stadium [41]. As a result, the following hypothesis was proposed:

H4. 

Effective crisis management is significantly impacted by passengers’ trust, resulting in increased satisfaction among commuters.

A service relationship’s effectiveness is greatly influenced by both customers and service providers [43]. Customers’ degree of agreement has a big impact on their perception of service quality and customer trust. Additionally, the distinctive and vital role of service provider orientation fosters trust and enhances service quality. According to Kospandani and Wahyudi [44], trust is the main cause of loyalty when service quality and qualities are improved. In their analysis of Indonesia’s electric railway system, the COVID-19 strategy successfully gained customers’ trust in public transportation by analyzing satisfaction, trust, and loyalty during the pandemic. Loyalty to public transit is influenced by user perceptions of significance, the quality of service, and interpersonal relationships [34]. The importance of public transportation is highlighted by the fact that the best measures of its quality are the willingness to continue using it, overall satisfaction, and the use of the service. As a result, the following hypothesis was proposed:

H5. 

Trust among passengers significantly impacts customer satisfaction.

2.3. The SERVQUAL RATER Model

The SERVQUAL model, established by Parasuraman et al. [9], is one of the most prominent approaches for measuring product or service quality. It demonstrates that service efficiency is satisfactory when contrasted with customer expectations. It also symbolizes the discrepancy between what customers want and what they think has been rendered [45]. Service quality stands as a multifaceted concept which precedes the attributes of empathy (EP), assurance (AS), responsiveness (RS), tangible goods (TG), and reliability (RL), according to Parasuraman et al. [7]. EP refers to the customized care and attention given to customers, whilst RS stands for passion and prompt responsiveness to consumer demands and difficulties [7]. In addition to expertise and courtesy, the word AS refers to an employee’s capacity to gain the trust and confidence of clients. RL is the capacity to accurately and consistently provide the promised service [7]. Additionally, TG includes the facilities, furnishings, and external appearance of the service providers [8]. Customer perception, expectation, satisfaction, or attitude can also be used to gauge the quality of a service [46].

Numerous research on the service industry has found AS to be one of the most substantial elements affecting service quality [47]. According to Chuenyindee et al. [8], it had a substantial impact on PUV service quality and customer satisfaction during the COVID-19 epidemic in the Philippines. This is because it makes it possible to evaluate how responsibly the transport company’s employees carry out their responsibilities regarding route knowledge and safety [48]. However, because AS indicates consumers’ confidence in the service provider, which is significant but ranked second to RL and RS, it was ranked third in the hierarchy of significance for evaluating service quality in the study by Knutson et al. [49]. Guests prioritize receiving consistent, dependable service and prompt attention, and once these needs are met, the confidence in staff competence becomes more relevant. Nonetheless, it still portrays a significant effect on commuter satisfaction.

RL was also found to be a significant element that influences service quality [8,50]. Dhaka City’s paratransit passengers reported that they are not as happy with their commute times on workdays, which has an impact on the paratransit’s reliability rating [48]. Providing effective training and management for the Doha Metro staff to enhance their responsiveness to customer needs and demands can significantly impact service quality. Abd-El-Salam et al. claims that RL influences customer satisfaction by affecting trust and a client’s overall view of a service after utilizing it [51].

Furthermore, the ability of service providers to empathize and connect with the people they serve has a significant impact on their connection [52]. A study by Alam et al. [53] found that EP had the most significant influence on customer satisfaction when assessing railway services. Additionally, clients perceive a link between their behavioral objectives and satisfaction once service providers demonstrate important levels of EP, as explained by Chuenyindee et al. [8]. Furthermore, results from a study of the literature by Cavana et al. [54] indicated that EP was one of the elements that significantly impacted the overall quality of the services.

The service business, on the other hand, depends heavily on TG, which encompasses all correspondence components, physical manifestation, and the individuals who provide services, including their outward appearance [7]. When assessing pedestrian amenities (tangibles), it is essential to consider the routes leading to and from the stations [55]. Alam and Mondal [53] evaluated the service offered by the railway of Khulna slums, emphasizing the necessity of cleanliness and visual attractiveness as tangible elements for reaching high customer satisfaction. In light of this, it can be concluded that TG is one of the most important variables affecting service quality, as indicated by the majority of SERVQUAL research [7,56,57,58,59].

Lee et al. [60] defined RS as the efficiency, speed, and customer acknowledgment of service. Chou et al. [61] stressed the critical role of RS in service quality, noting that a more responsive service provider leads to greater customer satisfaction. Additionally, Tiglao et al. [62] emphasized that the overall RS significantly influences how paratransit users in the Philippines perceive the quality of their services. Thus hypothesizing the following:

H6. 

Service quality domains greatly affect passenger satisfaction.

While the SET and SERVQUAL RATER models have been broadly used in public transportation research, their integration plus analysis with machine learning techniques is a relatively novel approach, especially in the context of Middle Eastern metro systems. This study builds on the work of Ong et al. [12] and Wang et al. [13] and applies a machine learning integrated algorithm to forecast passenger satisfaction with the Doha Metro, focusing on both the perceived trust (from SET) and service quality dimensions (from RATER SERVQUAL). This integration permits a more nuanced interpretation of how these factors interact to influence overall satisfaction.

In the case of the Doha Metro, this approach has the potential to uncover deeper insights into the drivers of passenger satisfaction, enabling more targeted improvements in service quality, using the established framework, concept, and findings from linked investigations. The integrated conceptual framework, created using the SERVQUAL RATER model plus the Social Exchange Theory (SET), is shown in Figure 1. Based on the earlier parts outlining the hypothesis-generating procedure, six hypotheses were created.

3. Methodology

3.1. Data Collection and Participants

This survey was intended for all expatriates in Qatar who have used Qatar’s autonomous train network. Through the use of purposive sampling, the online survey was disseminated through different types of platforms such as Google Forms, Facebook Messenger, and WhatsApp. In order to make sure that respondents were aware of and had utilized the Doha Metro system, this strategy sought to contact people who fit the precise requirements. To ensure that their answers appropriately represented their individual experiences and viewpoints on the metro service, participants were expected to finish the survey on their own. Informed written consent was acquired prior to distribution, and the university ethics committee board approved the study and the survey questionnaire. A 99.38% response rate was obtained from the analysis of 319 genuine responses out of 321 respondents.

According to Hair et al. [63], a sample size of 250 respondents is sufficient for generalization when using subjective measures. In this study, a purposive sampling approach was employed. As noted by Andrade [64], purposive sampling is effective for generalizing findings to the population or subpopulation from which the sample is drawn, ensuring representation of the target group. Given that less than 30% of the population has experience using the Doha Metro, purposive sampling was selected to focus specifically on individuals with firsthand experience. This approach enables the extraction of meaningful insights from those who have used the metro, allowing for a more accurate representation of the generalized public’s perception. To further enhance the robustness of our findings, we leveraged machine learning algorithms, specifically Long Short-Term Memory (LSTM), to forecast public sentiment and insights [65].

The demographic profile of the respondents was predominantly male, making up 70.22% (224 individuals) of the total sample, while females accounted for 29.78% (95 individuals). In many expatriate communities, particularly in Qatar, the workforce is predominantly male due to employment opportunities in industries like construction, engineering, and transportation [66], which attract more male workers. Regarding age distribution, the majority of the participants fell within the 26–35 age range (41.07%), followed closely by those aged 36–45 (40.44%). A smaller percentage of respondents was aged 46–55 (12.85%), while younger individuals aged 18–25 accounted for 3.44%. Likewise, respondents aged 56–65 made up only 2.19% of the sample. The working-age population dominated the expatriate community, as many foreign nationals have come to Qatar for employment [4]. That is, the Doha Metro is frequently used by professionals and daily commuters, who typically fall within this age range.

In terms of usage experience with the Doha Metro, none of the respondents reported having no prior experience with the transit system. Most respondents had been using the metro for 1 to 2 years (41.07%), while 38.87% had been regular users for three years. A smaller segment of the population (20.06%) had been using the metro for less than a year. The Doha Metro launched in 2019, so it is still relatively new [5]. The high percentage of users in the 1- to 3-year range suggests that many respondents adopted to the metro soon after its launch and continued to use it regularly.

A significant majority of respondents (82.76%) resided near a Doha Metro station, while 17.24% did not live in close proximity. Living near a metro station makes daily commuting more convenient, encouraging frequent use. Additionally, many residential areas and accommodations for expatriates are strategically located near metro stations. The most commonly mentioned nearest metro station was Al Mansoura (19.75%), followed by Al Messila (8.15%), Al Sadd (7.84%), and Al Aziziyah (7.21%). These stations are located in densely populated residential and commercial districts, making them highly accessible. Meanwhile, several stations had no respondents listing them as their closest station, such as Al Riffa, Al Shaqab, Katara, and the Doha Exhibition and Convention Centre. The lack of responses for certain stations suggests that fewer people live in those areas or that they are primarily commercial zones.

When asked about car ownership, 56.43% of respondents stated that they did not own a car, while 43.57% indicated that they had their own vehicle. Many expatriates in Qatar, especially those from middle-income backgrounds, rely on public transportation due to the high costs associated with car ownership, fuel, insurance, and maintenance [37]. That is, the Doha Metro provides a cost-effective alternative for daily commuting.

In terms of employment status, an overwhelming majority of respondents were employed (94.36%), while a smaller percentage were self-employed (2.19%), students (1.57%), or unemployed (1.88%). Since Qatar’s foreign nationals primarily migrate for work, the high employment rate aligns with expectations [67]. Students, self-employed individuals, and unemployed respondents form a small proportion of the total sample.

Moreover, most participants (62.70%) held an undergraduate degree, while 27.59% had completed a diploma course. A smaller percentage (7.21%) had obtained a master’s degree, and only 0.94% held a doctoral degree. A few respondents (1.57%) reported having no formal higher education. The educational distribution suggests that most metro users are professionals or skilled workers. The relatively low number of respondents with advanced degrees (master’s or doctoral) could indicate that higher-income professionals may prefer private transport.

The respondents reported varying monthly income levels, with the largest proportion (20.69%) earning between QAR 11,001 and QAR 14,000 (USD 3000–3800). Those earning QAR 8001 to QAR 11,000 (USD 2190–3000) made up 19.75%, while 18.81% earned between QAR 5001 and QAR 8000 (USD 1370–2190). A smaller portion (15.05%) was in the QAR 14,001 to QAR 17,000 (USD 3800–4600) range, while those earning above QAR 20,001 (USD 5400 and above) accounted for 5.96%. Additionally, 2.82% of respondents reported having no income due to unemployment. The Doha Metro is widely used by middle-income workers, which explains why most respondents fall within this salary range [6]. In accordance, high-income earners are more likely to own cars and rely less on public transportation.

When asked about their primary reason for using the Doha Metro, respondents provided a variety of responses. The most common reason was work-related travel (36.36%), followed by personal or essential trips (35.11%). Additionally, 25.39% of participants used the metro for leisure purposes, while school-related (1.25%) and family-related travel (1.88%) accounted for a smaller percentage of responses. The metro is primarily used for commuting to work, as many expatriates rely on public transportation for daily travel. The significant percentage of leisure-related trips suggests that the metro is also popular for visiting malls, parks, and entertainment areas on weekends [13].

3.2. Questionnaire and Data Preprocessing

There were three primary sections to the English-language survey. The purpose of Section 1 was to gather demographic data, such as the respondents’ gender, age, length of time utilizing the Doha Metro, place of residence, employment position, educational achievement, and main motivation for utilizing the autonomous train network. The Socio-Economic and Environmental Transportation (SET) framework’s indicators—safety (SA), accessibility (AC), perceived economic benefit (PEB), crisis management (CM), and trust (TR)—are the subject of Section 2. The SERVQUAL RATER dimensions—which were derived from the pertinent literature sources [8,9,68]—are used in Section 3 to quantify service quality. These dimensions included empathy (EM), responsiveness (RP), tangibles (TG), assurance (AS), reliability (RL), and service quality (SQ).

As shown in

Table 1. Construct and measurement items.

Variables	Code	Construct	References
Accessibility	AC-1	Doha Metro is accessible on weekdays.	[9,62]
	AC-2	Doha Metro is accessible on weekends/holidays.	[9,62]
	AC-3	Doha Metro is accessible throughout daytime business hours (6 a.m. to 6 p.m.).	[9,62,69]
	AC-4	Doha Metro is accessible outside of regular working hours (6 p.m. to 6 a.m.).	[9,62,69]
	AC-5	Waiting time at metro train station is short.	[9,62,70]
	AC-6	There are adequate trainsets.	[9,69]
	AC-7	Doha Metro train stations are in easily reachable locations.	[9]
	AC-8	Family and PWD sections are available inside metro trainsets.	[9]
	AC-9	There is dedicated bicycle and car parking for passengers of Doha Metro.	[71]
	AC-10	I appreciate Doha Metro’s free feeder bus service which offers transfer from a bus station to Doha Metro station.	[72]
Safety	SA-1	There is a specific order for boarding and alighting the metro train.	[9,73,74,75]
	SA-2	Accidents are uncommon when riding the metro train.	[9,35,76]
	SA-3	Crimes are uncommon on metro trains.	[9,35,76]
	SA-4	Safety measures are in place for both staff and customers.	[9,35,76]
	SA-5	I feel safe riding the metro train even if it is fully automated and without an on-train operator and staff.	[77]
	SA-6	I don’t mind sharing space with other riders on the metro train.	[9,71,74,78]
Perceived Economic Benefit	EB-1	For me, the Doha Metro fare/travel pass are fair and affordable.	[9,79,80,81]
	EB-2	Taking the metro rail allows me to save money.	[9,80,82]
	EB-3	I use the Doha Metro train due to its affordability.	[9,80,82]
	EB-4	I think using Doha Metro can contribute to reducing carbon emissions and road traffic congestion.	[83]
	EB-5	I believe using the metro train can increase economic growth.	[84,85]
	EB-6	I believe that transitioning from using private to public transportation has a positive economic impact.	[77]
Crisis Management	CM-1	Passengers of Doha Metro follow health and safety precautions, i.e., COVID-19 precautions.	[86]
	CM-2	Staff of Doha Metro follow health and safety precautions, i.e., COVID-19 precautions.	[86]
	CM-3	The government, in my opinion, cares about the health and safety of commuters.	[40]
	CM-4	I trust that the government ensures the implementation of appropriate health and safety measures in the Doha Metro system and stations.	[40]
	CM-5	Any health and safety issues are easily resolved as mandated by the government and Doha Metro.	[40]
Trust	TR-1	I think using public transportation such as Doha Metro would create employment opportunities.	[40]
	TR-2	I think that Doha Metro is an essential public transportation.	[78]
	TR-3	I feel comfortable riding the Doha Metro trains.	[35]
	TR-4	I trust Doha Metro for my daily commute.	[78]
Satisfaction	SAT-1	I find it easy to navigate within train stations to reach my destination.	[81]
	SAT-2	I did not encounter any technical issues during my travel using the Doha Metro.	[87]
	SAT-3	Payment methods (e.g., station travel card vending machine, website, and app) are working properly.
	SAT-4	Travel distance between stations is satisfactory.	[88]
	SAT-5	The quality of service provided by the Doha Metro is good.	[81]
	SAT-6	Doha Metro offers a safe and secure overall experience.	[81]
Assurance	AS-1	Doha Metro staff are well trained and can communicate effectively to a customer’s query.
	AS-2	Doha Metro staff are polite.	[81]
	AS-3	I feel safe when doing Doha Metro transactions.	[89]
	AS-4	Doha Metro staff are courteous to inform customers of the changes in the process.	[89]
	AS-5	The method of informing customers is appropriate when an abnormal situation happens.
Reliability	RL-1	Doha Metro is capable of delivering service that has less-to-no errors.	[90]
	RL-2	I think that Doha Metro is a reliable public transportation system.	[91]
	RL-3	Doha Metro provides the service accurately on the first attempt.	[92]
	RL-4	Staff of Doha Metro are trustworthy and dependable.	[91]
	RL-5	When customers have a problem in their transactions, Doha Metro staff show sincere interest in solving it.	[91]
	RL-6	Train arrival is on schedule without delays.	[75]
	RL-7	Reasonable time can be expected to complete the travel.	[75]
Empathy	EM-1	The passengers of Doha Metro are given individualized and personal attention.	[89]
	EM-2	Doha Metro employees understand specific needs and problems of the customers.	[89]
	EM-3	The customers of Doha Metro are treated importantly.	[89]
	EM-4	Doha Metro puts their customer’s interests and needs as a priority.	[89]
	EM-5	Doha Metro employees are courteous when communicating with customers.	[93]
Tangibles	TG-1	Doha Metro stations, trains, and feeder buses create an environment that makes passengers feel at ease and valued.	[94]
	TG-2	The station’s layout is strategically designed for maximum efficiency, ensuring smooth flow and ease of access.	[95]
	TG-3	Doha Metro stations are tidy and consistently maintained to high standards.	[87,88]
	TG-4	Doha Metro offers ample seating and standing space in each trainset.	[75,81]
Responsiveness	RP-1	The customers of Doha Metro are given individualized and personal attention.	[89]
	RP-2	Doha Metro staff/help desks are available whenever the customer needs them.	[91]
	RP-3	Doha Metro responds promptly and immediately to the customers.	[91]
	RP-4	Doha Metro is keen to provide answers to customer questions and problems.	[91]
	RP-5	Doha Metro staff are courteous when communicating with customers.	[93]

, the questionnaire was created following a thorough literature assessment to guarantee the suitability and reliability of the chosen indicators. A 5-point Likert scale was used to assess multiple constructs, and all measurement variables were derived from existing studies [65].

Data preprocessing was carried out before incorporating the machine learning algorithm (MLA). During the data-cleaning phase, correlation analysis was used, and indicators with p-values greater than 0.05 were removed due to their lack of statistical significance. Furthermore, for MLA optimization, only indicators with correlation coefficients greater than 0.20 were maintained. Tested both directly and indirectly, none of the indications were removed since all were significant, and these underwent data aggregation through mean calculation and served as the input variables needed for the machine learning algorithm [65].

3.3. Neural Network Algorithm

To evaluate and forecast passenger satisfaction with the Doha Metro based on aspects of service quality, this study used an artificial neural network. The fact that it is based on the functions of the human brain makes it perfect for behavior prediction [95]. With the growing global emphasis on e-public transportation, machine learning techniques offer an innovative approach to evaluating service quality, optimizing operations, and enhancing commuter experiences in autonomous train networks [12]. In transportation research, neural networks are a useful tool for assessing both structured and unstructured data because of their well-known capacity to simulate intricate, nonlinear interactions between variables [96]. The use of MATLAB with the Levenberg–Marquardt algorithm was considered in this study using MATLAB 2024b, leveraging its built-in machine learning and deep learning toolboxes to streamline model design, training, and evaluation. According to the findings of Ozturk and Basar [97], optimizing the number of nodes in the hidden layer will result in the most effective classification model.

The LSTM network was added to improve model performance and predict the generalized satisfaction. LSTM is a specialized type of recurrent neural network (RNN) that enables the model to capture patterns in passenger perceptions over time. This approach aligns with previous studies that have successfully used LSTM to forecast trends in urban rail transit ridership [12]. From this, it could be deduced that machine learning-based techniques offer useful insights for forecasting demand, boosting service reliability, and raising customer satisfaction [97], since e-public transportation systems, like the Doha Metro, depend on automation and real-time operational efficiency.

Through the integration of machine learning methodologies, specifically neural networks and LSTM, this study offered a data-driven method for evaluating transport service quality. As e-public transportation grows globally, the integration of machine learning in transit analysis has the potential to significantly improve efficiency, sustainability, and user experience [98,99]. The conclusion provides practical advice for increasing commuter satisfaction and supporting evidence-based decision-making for urban mobility planning in Qatar.

4. Results

4.1. Neural Network

A total of 30 iterations per parameter combination (70:15:15; 80:10:10; 90:5:5 for the training:testing:validation ratio) were performed with 10 and 20 hidden layers using the Levenberg–Marquart algorithm. The purpose of this approach was to identify the most stable output with the highest r-squared value, a key indicator of predictive accuracy. According to Öztürk and Başar [97], an r-squared value between 70% and 90% is considered acceptable, with higher values indicating greater predictive power. As shown in Figure 2, the analysis achieved a high r-squared value of 85.91% overall and 96.53% on the testing output, confirming the overall robustness of the model and prediction of satisfaction. This was further supported by the 86.42% validation output from the 84.51% training output.

The best-performing combination was identified as the 80:10:10 with 20 hidden nodes, as this configuration consistently produced high r-squared values across multiple iterations. Additionally, this model required fewer epochs to converge, reducing computational time while maintaining accuracy. Further optimization was performed to evaluate error rates for each additional hidden layer. The results suggest that the 80:10:10 configuration with 20 hidden layers can serve as an effective decision support system for predicting passenger satisfaction in e-public transportation systems like the Doha Metro.

4.2. Long Short-Term Memory (LSTM)

The optimum parameter served as the processing for completing the LSTM optimization process, which only had 1 min and 2 s over 250 epochs of a complete iterative run. The model’s initial observed output (Figure 3) demonstrated a low root mean square error (RMSE) of 0.5054, indicating a reasonable predictive performance. Further model updates and LSTM tuning improved accuracy, yielding an even lower RMSE of 0.0817 (Figure 4). This significant reduction in error highlights the enhanced forecast capability of the LSTM model after optimization.

Figure 5 illustrates the forecasted passenger satisfaction regarding e-public transportation (Doha Metro). The y-axis represents the satisfaction (one as strongly dissatisfied and five—strongly satisfied), while the x-axis represents the responses. The results revealed a predominantly positive sentiment, with satisfaction scores clustering between four and five on the scale. This suggests that most passengers are either satisfied or very satisfied with the service provided by the Doha Metro. The high satisfaction levels among the Doha Metro commuters can be attributed to its exceptional service quality, reliability, safety, and accessibility. The metro maintains a service reliability of 99.85% and a punctuality of 99.66%, ensuring minimal delays and a seamless travel experience [100]. Safety remains a top priority, with an accident frequency rate of just 0.01 and approximately 90% of passengers perceiving it as a secure mode of transport. Furthermore, customer satisfaction surveys report an overall approval rating of 99.66%, reflecting the metro’s commitment to service excellence. The system’s accessibility, air-conditioned trains, comfortable seating, and integration with other transport options further enhance commuter convenience [6]. These factors collectively justify why the majority of passengers rate their experience positively, with satisfaction scores clustering between four and five.

Since predictive power was high (total r-squared = 85.9% for ANN) and the root mean square error was low (RMSE = 0.0817 for LSTM), a validation test was conducted using the SHapley Additive exPlanations (SHAP) package to determine the score of importance among the variables [97].

Table 2. Normalized score of importance.

Variable	Code	Importance	Normalized Importance (NI)
Empathy	EM	0.125	100.0%
Safety	SA	0.124	99.9%
Reliability	RL	0.113	90.8%
Trust	TR	0.113	90.8%
Responsiveness	RP	0.101	81.0%
Crisis Management	CM	0.097	77.6%
Economic Benefit	EB	0.090	72.6%
Assurance	AS	0.090	72.5%
Tangibles	TG	0.082	66.0%
Accessibility	AC	0.065	52.0%

presents the normalized importance scores of the key service quality variables influencing passenger satisfaction. Moreover, further classification techniques using different decision tree algorithms were employed. As depicted in the study of Chen et al. [101], decision trees provide significant factors that would delineate the target objective. However, only those which are highly significant would be present in the output. In this case, the random forest classifier optimum output was presented (Figure 6).

With the optimum parameters of depth 6, 80:20 training and testing ratios, and gini and best as the criterion and splitter, as well as a 0.1 learning rate, an average of 93% accuracy was obtained. This was run with 100 iterations together with the other combinations. With similar parameters, the basic decision tree was considered, but it only obtained a score of 88% accuracy rate on average. This is because Chen et al. [101] and others [18,65] depicted that random forest has the capability to identify the optimum output every iteration and present this as a result as compared to the basic decision tree. This highlights how the basic decision tree only provides the first output every iteration. In accordance, other decision tree algorithms were tested like LightGBM, XGBoost, and CatBoost. From the results, LightGBM accounted for only 81.255% while XGBoost resulted in 83.73%, and CatBoost resulted in an accuracy of only 17.5% with scattered predictions. This is because the decision tree algorithm and its corresponding parameters are also dependent on the dataset available and thus need to be tested across different algorithms to depict the best and optimum output [18,65]. Similarly to the results of the study from Ong et al. [65], random forest also outweighed the other decision tree algorithms in comparison when dealing with consumer behavior. Summarized in

Table 3. Summarized decision tree algorithm output.

Algorithm	Accuracy Score in Percentage
Random Forest Classifier	93.00
Basic Decision Tree	88.00
XGBoost	83.73
LightGBM	81.22
CatBoost	17.50

are the outputs of the different decision tree algorithms.

From the random forest output, similar connotations of the important factors were seen. This resulted in the best output, which could depict that empathy (X₁) influences the overall satisfaction to which this parent node would consider safety (X₀) with higher X₀, leading to a positive reliability (X₂) and trust (X₄), both of which would end up with high and very high satisfaction once both are received by the customer. On the other hand, responsiveness (X₃) will only be achieved when X₁, X₂, and X₄ are high. This means that only when customers feel that there is empathy, when the services could be trusted and reliable, would they feel that the overall service is responsive.

5. Discussion

The variable importance analysis identified the key factors influencing passenger satisfaction with the Doha Metro’s e-public transportation system. The study ranked the variables based on their predictive importance using the SHAP package. The highest importance was seen in empathy (EM) and safety (SA), followed by reliability (RL), trust (TR), responsiveness (RP), crisis management (CM), economic benefit (EB), assurance (AS), and tangibles (TG). In contrast, accessibility (AC) (52.0%) has a lower impact on satisfaction in this context, which is found to be below the considered threshold of 60%. This threshold was based on the findings of German et al. [96] who implied that SHAP output should be greater than 60% for it to be considered measurably significant. The results aligned with previous research that emphasizes the importance of service quality dimensions in shaping public transit user perceptions [7,8]. The findings provide critical insights into which factors should be prioritized to enhance the Doha Metro’s service quality and commuter experience.

The results showed that the two most influential criteria were EM (100%) and SA (99.9%), implying that passengers place a high importance on personalized service and security measures when utilizing public transportation. EM was the strongest predictor, supporting previous research findings that passenger interactions with staff, clear communication, and staff responsiveness significantly improve satisfaction [61]. Public transportation users, particularly in an autonomous train network like the Doha Metro, may continue to value human interaction for support, advice, and problem resolution despite its automation [55]. Strengthening customer service initiatives prior to boarding the metro, maintaining station staff member availability, and increasing real-time passenger support could all help to improve the commuter experience. Contrary to the notion that commuters would consider there to be no contact due to automation [102], it could be said that social and individual factors should still be considered. As a reflection, there are still people that would need assistance, support, and overall service for satisfaction to be enhanced—leading to a positive experience in using the Doha Metro.

SA had almost as much of an impact as EM, highlighting the significance of security, emergency procedures, and passenger trust in the system’s safety measures. This outcome is in line with earlier research showing that perceived safety has a significant impact on the use of public transportation [33]. As a fully automated, driverless system, the Doha Metro can increase passenger trust and encourage long-term ridership by upholding a high standard of safety awareness through well-lit stations, real-time monitoring, and efficient incident response procedures [11]. According to several studies, the public’s acceptance of automated transportation is not usually driven by safety concerns. Kim and Horrey [103] found that trust in technology and perceived control were more important than safety concerns in the acceptance of self-driving automobiles [103]. This could be attributed to participants’ familiarity with automation, which leads to increased baseline trust. In contrast, the present study on the Doha Metro emphasizes safety as a critical aspect because of its completely automated, driverless design, which gives passengers less direct control. Features such as real-time monitoring and emergency procedures increase trust, making public transportation safer than personal autonomous vehicles.

Additionally, both TR (90.8%) and RL (90.8%) were highly significant, confirming that passengers build trust based on their experiences and anticipate reliable, consistent service. In the context of public transportation, RL is defined as the minimal disruptions experienced by commuters, service availability, and on-time performance. According to research, commuters are more inclined to routinely use public transportation when they believe it to be dependable [13]. This impression is probably influenced by the Doha Metro’s high-frequency service and effective operations, but sustained efforts to reduce delays and improve system reliability will be essential for long-term acceptance. TR, on the other hand, was as significant, emphasizing that views of system fairness, transparency, and responsiveness to issues are all aspects of passenger confidence in the metro system that go beyond operational reliability [78]. That is, trust and satisfaction can be further increased by enhancing public communication, upholding service transparency, and skillfully responding to commuter comments.

RP (81.0%), CM (77.6%), and EB (72.6%) were the next set of key elements that highlighted how system responsiveness, affordability, and crisis readiness shape commuter experiences. RP measures how well the metro system and its employees address questions, complaints, or problems raised by passengers. Passenger satisfaction may be increased by prompt technical problem solving, efficient digital support systems, and real-time service updates because the Doha Metro is automated [82]. For CM, highlighting the importance of preparedness in handling emergencies or disruptions could positively impact commuter satisfaction. Previous studies have shown that public transit systems that implement clear, well-communicated crisis responses—such as health and safety measures—tend to retain higher passenger confidence and trust [9]. In accordance, EB suggests that affordability plays a role in passenger decisions, even in a high-income country like Qatar, where private vehicle ownership is high [37]. Pricing strategies, subsidies, and travel pass programs could enhance perceived economic benefits and encourage greater ridership. While the current study stresses the relevance of responsiveness, crisis management, and perceived economic benefit in automated transportation systems, previous research has revealed that these elements may not always be as important. For example, a study on autonomous car acceptance discovered that characteristics such as perceived control and faith in technology had a greater impact than crisis management or responsiveness [104]. Furthermore, in Swedish research on driverless buses, riders stressed convenience and accessibility over perceived economic benefit [104]. These variations may develop because passengers in automated systems, such as the Doha Metro, where human interaction is limited, place a higher importance on crisis management and responsiveness because they feel less in control of the system. In contrast, traditional or semi-automated transportation systems may not raise the same issues, enabling other variables to take precedence.

Additionally, both AS (72.5%) and TG (66.0%) were above the significance level, indicating that they have a significant impact on passenger satisfaction. AS has to do with how competent, polite, and professional the metro’s employees and system administration are seen to be. Despite being automated, passenger satisfaction can be increased by making sure support personnel and digital platforms provide clear and assured instructions [78]. Additionally noteworthy was the TG, which comprise the physical surroundings, sanitation, station esthetics, and infrastructural quality. Research indicates that well-maintained transit surroundings contribute to good evaluations of service quality; however, their influence is not as great as that of service-related elements [12]. According to some studies, tangibles like station esthetics and infrastructure quality may have a smaller impact on customer satisfaction than operational variables. Soza-Parra et al. [105] found that service reliability, namely headway regularity and crowding, has a greater impact on user satisfaction than physical environment attributes [105]. However, maintaining high TG is still necessary, as well-maintained transit areas contribute to positive impressions of the service quality. To maintain high standards, transportation authorities can implement regular maintenance and cleaning schedules, invest in infrastructure upgrades, improve station esthetics, and set up passenger feedback mechanisms. These strategies can enhance the commuter experience while reinforcing good passenger perceptions.

In this study, accessibility (AC) was found to have a lower impact (52.0%) on passenger satisfaction, falling below the considered significance threshold. While this suggests that accessibility may not be a primary factor influencing satisfaction, it does not imply that it is unimportant. Instead, it indicates that commuter expectations regarding accessibility are largely being met by the Doha Metro’s existing infrastructure, including station locations, feeder bus services, and overall accessibility features. However, this finding warrants further investigation, as accessibility is widely recognized as a key determinant of public transport usage. The perceived insignificance of accessibility may stem from survey design limitations or the underrepresentation of certain commuter groups, such as persons with disabilities (PWDs), elderly passengers, or those in areas with limited connectivity. Future research should address these gaps by refining survey questions, incorporating qualitative insights (e.g., focus groups or interviews), and re-examining accessibility indicators to capture a more diverse range of commuter experiences.

To further improve accessibility perceptions, enhancements should focus on first- and last-mile (FM/LM) connectivity, which plays a critical role in transit adoption. Research suggests that integrating shared mobility services, such as bike-sharing and on-demand shuttles, can help commuters travel between metro stations and their destinations more efficiently [106]. Additionally, improving pedestrian infrastructure can enhance walkability and overall commuter convenience [107]. Introducing On-Demand Multimodal Transit Systems (ODMTS), as seen in pilot programs in cities like Atlanta, could further reduce the reliance on private vehicles and promote greater public transport usage [108]. While the results indicate a lower direct impact of accessibility on passenger satisfaction, enhancing connectivity, inclusivity, and multimodal integration remains essential for optimizing metro ridership and improving the overall commuter experience.

5.1. Theoretical Contribution

This study significantly enhanced the existing body of knowledge on sustainable e-public transportation by integrating the Social Exchange Theory (SET) and the SERVQUAL RATER model with advanced machine learning methodologies. While previous research has largely depended on traditional multivariate analyses like structural equation modeling (SEM), this study pioneered the use of neural network and Long Short-Term Memory (LSTM) models to predict passenger satisfaction with greater accuracy and efficiency. This innovative approach enabled a deeper exploration of service quality dimensions and their complex effects on commuter perceptions [65]. Additionally, by concentrating on the Doha Metro, this study provided valuable insights into urban e-mobility in emerging economies with unique socio-economic and cultural factors. The research highlights the relevance of SET in understanding trust, reliability, and economic benefits in e-public transit while showcasing the potential of machine learning to improve transport service assessment and strategic planning [13]. Furthermore, this study emphasized the potential for scalable and adaptable methodologies that can be implemented in similar transit systems globally.

By combining machine learning methods like neural network and Long Short-Term Memory (LSTM) models with the Social Exchange Theory (SET) framework and the SERVQUAL RATER framework, this research improved the understanding of sustainable e-public transportation in comparison to other well-established models. Unlike traditional methods such as structural equation modeling (SEM), which requires manual intervention for model fitting, this approach offers greater accuracy and efficiency in predicting passenger satisfaction. Furthermore, this study acknowledges the methodological trade-offs between machine learning techniques and traditional statistical models like SEM. While SEM is valued for its ability to establish causal relationships, its application in large, complex models may lead to model misspecifications that require additional analyses [14,15].

In contrast, LSTM and neural network models offer enhanced predictive performance, particularly in handling nonlinear relationships and large datasets [17,18]. However, to address concerns regarding interpretability, this study utilized SHapley Additive exPlanations (SHAP) to identify key contributing factors, thus improving model transparency. The findings highlight the potential of hybrid modeling approaches that integrate SEM’s theoretical rigor with machine learning’s predictive strength, offering a more comprehensive assessment of service quality and passenger satisfaction. Since the assessment focused solely on the Doha Metro, future studies can be conducted once data are collected from various countries such as the United States, China, Mexico, Egypt, South Korea, and the Netherlands, to name a few. This would enable a comparative analysis of passengers’ perceptions of service quality, trust, and satisfaction in e-public transportation across both developing and developed countries. Additionally, forecasted behavior, as examined in this study, could be compared to assess how Qatar aligns with other nations. Such an approach would provide a more comprehensive and generalized understanding of e-public transportation utilization, service quality, and passenger satisfaction on a global scale. Lastly, the novel integrated model provided higher accuracy for forecasting. This implies that this could be used to assess and identify significant service-related factors affecting the overall satisfaction of e-public transportation.

5.2. Practical and Managerial Implications

The findings offered concrete, actionable recommendations for policymakers, urban planners, and public transportation authorities striving to improve service quality and commuter satisfaction in e-public transit networks. This study highlighted empathy, safety, and reliability as critical determinants of passenger trust and satisfaction. As a result, transit authorities should invest in extensive customer service training programs, reinforce security measures, and adopt predictive maintenance solutions to ensure operational reliability through real-time monitoring and diagnostics [8]. Additionally, improving accessibility through integrated multimodal transportation solutions and enhancing first- and last-mile connectivity could significantly boost ridership. Key strategies include adding e-scooter and bike-sharing stations close to metro hubs, increasing feeder bus lines in accordance with train schedules, and collaborating with ride-sharing services for affordable short trips. First- and last-mile connectivity can also be enhanced by park-and-ride locations, smart navigation apps, and pedestrian-friendly infrastructure. By making public transportation smoother, these policies lessen the dependency on private automobiles and promote increased commuter use.

Another key managerial implication is the necessity of crisis management preparedness, particularly in addressing unexpected disruptions such as pandemics or large-scale events [21]. Effective crisis response strategies, including clear communication, rapid adaptation to new safety protocols, and contingency planning, can improve passenger confidence and service resilience. Moreover, affordability remains a major driver of ridership; implementing strategies such as dynamic pricing models, government subsidies, and corporate partnerships for employee transit programs could help encourage the broader adoption of e-public transport services [37]. The Doha Metro may improve crisis management and affordability by implementing clear communication mechanisms, such as mobile alerts, and respond quickly to safety standards during interruptions like pandemics or large-scale events. Examples include contactless ticketing and improved sanitization. Dynamic pricing mechanisms, government subsidies for vulnerable groups, and corporate partnerships for employee transit programs can all help to increase public transportation affordability. These solutions would boost passenger confidence amid the crisis and promote broader, more sustainable ridership. The insights gained from this research can also inform policy formulation aimed at reducing congestion and environmental impact while increasing public transport uptake in urban centers. Introducing ecologically friendly options, such as electric buses and bike-sharing programs, can help to cut emissions and promote eco-friendly transport. Offering reduced passes to frequent customers and encouraging Transit-Oriented Development (TOD) near metro stations can help make public transportation more inexpensive and convenient. These initiatives help to reduce congestion, environmental effects, and dependency on cars, thereby encouraging more people to use public transit.

Several initiatives can be implemented to increase accessibility for the Doha Metro, including improving services during non-peak hours (6 p.m. to 6 a.m.), lowering wait times by boosting train frequency, and assuring suitable trainsets with accessible amenities for families and people with disabilities. Expanding station accessibility through better walking paths, incorporating micro-mobility choices such as bikes and e-scooters, and improving parking facilities will all help to improve first- and last-mile connectivity [55]. Introducing On-Demand Multimodal Transit Systems (ODMTS) could minimize the dependency on private vehicles while increasing public transportation use. Furthermore, expanding and boosting the frequency of the free feeder bus service would make it easier for passengers to reach metro stations. These enhancements would increase general accessibility and convenience for passengers.

To effectively implement the proposed improvements in service quality, security, and pricing, a structured approach tailored to Qatar’s socio-economic and regulatory framework is essential. First, policymakers should collaborate with relevant government agencies to establish clear guidelines for service quality enhancements, including mandatory security reinforcements, digital monitoring systems, and predictive maintenance strategies. Additionally, pricing adjustments should be carefully aligned with national transportation policies, incorporating dynamic pricing models and subsidies for lower-income commuters to promote affordability without compromising revenue sustainability. The regulatory feasibility of these initiatives depends on integrating them into Qatar’s National Vision 2030, leveraging existing public–private partnerships for funding and operational support. Stakeholder engagement, including transport authorities, urban planners, and technology firms, is crucial to ensuring a seamless implementation process. Finally, continuous assessment through commuter feedback and real-time analytics will help policymakers refine these strategies, ensuring they remain responsive to evolving urban mobility needs.

6. Conclusions

This study presented a comprehensive evaluation of service quality and passenger satisfaction in sustainable e-public transportation, using the Doha Metro as a case study. By integrating SET, SERVQUAL, and machine learning methodologies, the research identified key determinants of commuter satisfaction and provided data-driven strategies for enhancing service quality. The findings emphasized the crucial role of empathy (NI = 100.0%), safety (NI = 99.9%), and reliability (NI = 90.8%) in fostering passenger trust (NI = 90.8%) and satisfaction while also underscoring the importance of accessibility (NI = 52.0%), economic benefits (NI = 72.6%), and crisis management (NI = 77.6%) in sustaining ridership [7].

As urban centers worldwide continue to invest in automated and electric public transportation, the insights from this study offer practical guidance for improving service delivery, optimizing transit operations, and promoting sustainable urban mobility. The results aligned with global sustainability goals, such as reducing carbon emissions, mitigating traffic congestion, and enhancing the livability of cities. Moreover, leveraging machine learning in public transportation analysis presents opportunities for continuous improvement through real-time monitoring, predictive analytics, and adaptive service enhancements. Policymakers and transit authorities can use these findings to shape future initiatives that align with smart city objectives and environmental sustainability. By addressing the identified gaps and implementing the recommended strategies, urban transit systems can become more efficient, inclusive, and sustainable in the long run.

Limitations and Future Research

Despite its contributions, this study has several limitations. First, it focuses solely on the Doha Metro, which may limit the generalizability of findings to other transit systems with different socio-economic and cultural conditions. While this study provides valuable insights into commuter satisfaction and service quality within the Doha Metro, its findings may have limited generalizability to other cities with different transit systems, cultural contexts, and urban mobility challenges. The unique socio-economic conditions, commuter demographics, and infrastructure in Doha may not fully reflect the experiences of metro users in other regions. To address this limitation, future research could conduct a multi-city comparison to examine how service quality factors influence commuter satisfaction in diverse urban transit environments. Moreover, analyzing metro systems in varied urban contexts—such as those in Asia, Europe, and North America—would provide broader perspectives on service quality and commuter satisfaction. This approach would help identify global best practices and context-specific improvements that could enhance transit system adoption.

Additionally, incorporating real-time operational data and commuter feedback from various cities could strengthen its applicability and impact, ensuring that its insights contribute to the advancement of sustainable e-public transportation on a global scale. To add, while machine learning models such as neural networks and LSTM demonstrated high predictive accuracy, their limited interpretability remains a challenge for users with minimal adoption. Although this study mitigated this issue by employing SHapley Additive exPlanations (SHAP) to enhance transparency, future research could further explore explainable AI (XAI) techniques, such as Local Interpretable Model-Agnostic Explanations (LIME) or hybrid SEM-ML models. Integrating SEM with machine learning techniques may provide a balanced approach that leverages the causal inference capabilities of SEM while maintaining the superior predictive power of machine learning. This approach could offer deeper insights into service quality determinants while ensuring methodological robustness. To add, future research may opt to consider demographic profiling as part of the effects among passenger behavior. This may result in more correlation among the socio-demographic profiling of respondents.

Moreover, while this study relied on survey-based data, future research should integrate real-time operational data (e.g., ticketing records, GPS tracking) to enhance the validity of the findings. A longitudinal study could also provide insights into evolving commuter preferences and the dynamic impact of service quality improvements over time. Future research may also consider other variables as extensions such as fuel prices, urban development projects, and infrastructures, which may also affect the overall satisfaction of passengers. In accordance, another limitation concerns the sample size. While purposive sampling was utilized to ensure that insights were drawn from actual Doha Metro users, future research should aim to increase the sample size to enhance the reliability and robustness of the results. Expanding the dataset with more survey responses or integrating additional data sources, such as user-generated feedback from digital platforms, could provide a more holistic representation of commuter sentiment.

Furthermore, this study may be subject to respondent biases and variations in self-reported experiences. Integrating real-time operational data, such as ridership patterns, trends, and commuter feedback from digital platforms or interviews and group discussions could enhance the robustness and applicability of the analysis. Another limitation lies in the dynamic nature of urban transit, where commuter satisfaction can be influenced by rapidly changing factors such as infrastructure developments, policy changes, and emerging technologies. Future studies should adopt longitudinal research designs to capture temporal trends and evolving commuter expectations.

Finally, this research does not thoroughly examine the cultural and behavioral elements impacting transit adoption, citing the high percentage of automobile ownership and the expatriate-dominated population of Qatar as obstacles to metro usage. To fill in these gaps, qualitative findings could be incorporated into future research to delineate changing behaviors among commuters.