Short-Term Traffic Congestion Prediction Using Hybrid Deep Learning Technique

Round 1

Reviewer 1 Report (Previous Reviewer 2)

The manuscript sent to me for review is interesting from a methodological and empirical point of view. The method of research and the results do not raise any doubts. The authors clearly described how they obtained the data, which is an added advantage. They took up quite an interesting topic which is traffic jams.

 

Conclusions were formulated correctly. The selection of literature is correct, although it could be extended.

 

Section headings should be on one page with the first paragraph of that section. They are currently on the previous page in isolation from the main text.

 

All statements are missing.

Author Response

Point 1: Conclusions were formulated correctly. The selection of literature is correct, although it could be extended..

 

Response 1: Yes, incorporated and extended the literature section.

 

1. Gollapalli et al. [25] proposed a cloud-based intelligent road traffic congestion prediction model powered by a hybrid neuro-fuzzy method to decrease vehicle delays at various road junctions. The proposed model aims to assist automated traffic management systems in lessening congestion, especially in smart cities where IoT sensors are deployed along the route.

 

Liu L et al. [26] devised a traffic congestion situation assessments (TCSA) strategy for a fuzzy integrated multi-metric assessment based on three predicted vehicle traffic variables aimed at the 5G Internet of Vehicles setting. The smoothing coefficient and model weight may be adjusted with the prediction algorithm. The trapezoidal membership function is used to calculate the traffic congestion index membership degree, and the adaptive CRITIC method is used to calculate traffic congestion evaluation metrics weights.

 

Finally, Bokaba T et al. [27] used data collected from a roadway in Gauteng Province, South Africa, to determine vehicle traffic flow in advance. In general, ensemble learning can improve the performance of weak classifiers. In this study, to compare the predictive performance of the proposed methods, a real-world dataset was used, as well as bagging, boosting, stacking, and random forest ensemble models. The ensemble prediction model was designed to forecast traffic congestion on roads.

 

Point 2: Section headings should be on one page with the first paragraph of that section. They are currently on the previous page in isolation from the main text.

 

Response 2: Yes, incorporated all section headings on one page with the first paragraph of that section.

Author Response File: Author Response.pdf

Reviewer 2 Report (Previous Reviewer 1)

Authors have addressed all my comments.

Author Response

Response to Reviewer 2 Comments

 

Point 1: Authors have addressed all my comments.

 

Response 1:

Reviewer 3 Report (New Reviewer)

The paper proposed a hybrid Xception- Support Vector Machine (XPSVM) classifier model to predict Short-Term Traffic Congestion. This study is very interesting. However, this study has some technical problems. Some comments that may help improve the quality of the study.

（1）      The depth and scope of section 2 (Literature Review) are inadequate. Besides, the structure of this section is also not reasonable enough. The related work is not only the introduction of the methods, but also show the readers the main background of this research field. So that the readers can easier to follow the manuscript and comprehend the motivation of it.

（2）      The depth of section 5 (Results) are inadequate. XPSVM needs to be compared with existing SOAT baseline methods to reflect the advantages of the XPSVM model. If possible, the authors are advised to add a map visualization that illustrates the spatial distribution prediction errors (maybe with other baselines), to showcase the model’s improved ability.

（3）      The authors are advised to provide at least information about experiment platform (including both hardware and software environment) and hyperparameter settings, especially when it comes to comparison between different baselines.

（4）      Typesetting issues hindered the readability of this manuscript. I think a manuscript with professional typesetting is respect for research. In addition, there are some typos and mistakes, the authors should polish their paper carefully.

（5）      Authors should refer to recent references, e.g.

[1] Urban traffic flow prediction: a dynamic temporal graph network considering missing values (doi: 10.1080/13658816.2022.2146120)

Author Response

Response to Reviewer 3 Comments

 

Point 1: The depth and scope of section 2 (Literature Review) are inadequate. Besides, the structure of this section is also not reasonable enough. The related work is not only the introduction of the methods, but also show the readers the main background of this research field. So that the readers can easier to follow the manuscript and comprehend the motivation of it..

 

Response 1: Yes, incorporated the depth and scope of Literature Review and the main background of this research field.

Road traffic congestion (RTCs) is a significant worldwide issue; when there is an increase in the demand for transit operations above the capacity of the system, unexpected capacity constraints (RTC) arise. Traffic congestion costs the South African economy billions of rands (ZAR) each year. On a worldwide basis, cities such as Bengaluru, India, are among the most affected by this issue. The top five most congested cities in the world are Istanbul (Turkey), Moscow (Russia), Kyiv (Ukraine), Bogota (Colombia), and Mumbai (India). Escalating traffic congestion is the direct and indirect consequence of a significant portion of road traffic collisions, leading to increasing number of accidents and mortality rates on highways worldwide [6].

 

The World Health Organization (WHO) confirmed that RTC causes health problems that lead to about 3.7 million deaths around the world, especially in developing cities, where money losses, delays, fuel waste, road accidents, and pollution are significant. In addition, sulfur oxide, carbon monoxide, and nitrogen oxide are produced. Moreover, according to the report, the primary pollutants in the air are connected to road traffic congestion, which causes ailments affecting the cardiovascular system. Congested roads pose a threat to those stuck in congestion as well as residents who live in close proximity to motorways [7].

 

In this research paper, an algorithm for forecasting gridlock on highways with supervised hybrid approaches is proposed. In deep learning, hybrid techniques enhance several classifiers and improve their effectiveness. As a result, hybrid methods have demonstrated superior efficiency in road traffic congestion applications [8].

 

Point 2: The depth of section 5 (Results) are inadequate. XPSVM needs to be compared with existing SOAT baseline methods to reflect the advantages of the XPSVM model. If possible, the authors are advised to add a map visualization that illustrates the spatial distribution prediction errors (maybe with other baselines), to showcase the model’s improved ability.

 

Response 2: Yes, incorporated. We included an XPSVM comparison and added a map visualization.

 

In below Figure, one can also notice the there were two or more instances in which the predicted value needed to be matched to the actual value, which mainly happened repeatedly when applying proposed algorithm but not when applying comparing algorithms. Therefore, we concluded that the existing algorithms did not perform better than the proposed model. The benefits of the proposed algorithm are listed below.

 

The benefits of the proposed algorithm are listed below.

• The proposed model best fits to linear and nonlinear map images because of its optimizer, the SVM classifier, which uses the L2 regularization technique, which is not present in the compared algorithms, and it also outperforms other algorithms in terms of training speed, fewer parameters (less memory con-sumption), weight sharing, and error rates. 

 

Map Visualization:

(a)	(b)

 

Figure 5. (a). Map with moderate congestion. (b). Map with High congestion

Above, Figure 5-(a) and (b) show the image maps that were used as input to the proposed algorithm to predict traffic congestion.

 

Point 3: The authors are advised to provide at least information about experiment platform (including both hardware and software environment) and hyperparameter settings, especially when it comes to comparison between different baselines.

 

Response 3: Yes, incorporated experimental platform and the hyperparameter settings.

 

The experimental platforms used for this research study are as follows:

• Hardware: i7 Processor, 16GB RAM, Graphics processing unit (GPU).
• Software: macOS Catalina, Python 3.10.5, API-Keras.

 

Hyperparameter settings:

In this study, we used the GridsearchCV hyperparameter tuner, as shown below in Table 1.

Hyperparameters	Input values	Best parameter values
epochs	10, 15, 20, 25, 30, 35, 40	25
Hidden Layers	14, 28, 42	14
neurons in hidden layer	10, 15, 20, 25, 30	10
dense layers	3, 6, 9, 12, 15	3
1st dense layer neurons	200, 300, 400	200
2nd dense layer neurons	100, 200, 300	100
3rd dense layer neurons	50, 100, 150	50
Dense layers activation functions	relu, sigmoid, softmax, softplus, tanh, exponential	relu
Kernel regularizer	L1,L2	L1
L2 regularizer weights	0.1, 0.01, 0.001	0.001
Loss functions	categorical hinge, categorical, categorical cross entropy	categorical hinge
Optimizer	adam, adadelta, sgd, adamax, nadam	adam

 

 

 

Point 4: Typesetting issues hindered the readability of this manuscript. I think a manuscript with professional typesetting is respect for research. In addition, there are some typos and mistakes, the authors should polish their paper carefully.

 

Response 4: Yes, incorporated all typesetting issues,typos and mistakes and polished the paper.

 

Point 5: Authors should refer to recent references, e.g.

 

[1] Urban traffic flow prediction: a dynamic temporal graph network considering missing values (doi: 10.1080/13658816.2022.2146120).

 

Response 5: Yes, incorporated and referred to recent references.

 

1. Gollapalli, M.; Musleh, D.; Ibrahim, N.; Khan, M.A.; Abbas, S.; Atta, A.; Khan, M.A.; Farooqui, M.; Iqbal, T.; Ahmed, M.S; Ahmed, M.I.B. A neuro-fuzzy approach to road traffic congestion prediction. Computers, Materials & Continua 2022, 73, 295–310.
2. Liu, L.; Lian, M.; Lu, C.; Zhang, S.; Liu, R; Xiong, N.N. TCSA: A Traffic Congestion Situation Assessment Scheme Based on Multi-Index Fuzzy Comprehensive Evaluation in 5G-IoV. Electronics 2022, 11, 1032.
3. Bokaba, T.; Doorsamy, W; Paul, B.S. A Comparative Study of Ensemble Models for Predicting Road Traffic Congestion. Applied Sciences 2022, 12, 1337.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report (New Reviewer)

Most of the concerns in the last-round of review have been addressed. However, the authors need to discuss the differences between this manuscript and the following work

WANG P X, ZHANG Y, HU T, el at. Urban traffic flow prediction: a dynamic temporal graph network considering missing values[J].International Journal of Geographical Information Science. 2022. doi: 10.1080/13658816.2022.2146120

Author Response

Point 1: Most of the concerns in the last-round of review have been addressed. However, the authors need to discuss the differences between this manuscript and the following work.

WANG P X, ZHANG Y, HU T, el at. Urban traffic flow prediction: a dynamic temporal graph network considering missing values[J].International Journal of Geographical Information Science. 2022. doi: 10.1080/13658816.2022.2146120

 

·         Response 1: Yes, incorporated the major differences between the manuscript and the given work, which are listed below.

 

manuscript (Short-Term Traffic Congestion Prediction using Hybrid Deep Learning Technique)	Work(Urban traffic flow prediction: a dynamic temporal graph network considering missing values)
This work predicts short term traffic congestion.	This work predicts urban traffic flow.
This work proposes the Xception support vector machine (XPSVM) model.	This work proposes a dynamic temporal graph neural network considering missing values (D-TGNM) model.
The proposed model was validated on an actual traffic dataset collected from Google Maps of Bangalore, Karnataka, Inida.	The proposed model was validated on an actual traffic dataset collected in Wuhan.
The proposed model forecasted traffic congestion with an accuracy of 97.16%.	The proposed model predicted results for four missing data scenarios (15% random missing, 15% block missing, 30% random missing, and 30% block missing).

 

Point 2: Moderate English changes required.

 

Response 2: Yes, incorporated the manuscript was edited by mdpi English editing services. However, the manuscript was also edited and polished, as evidenced by the English editing certificate shown below.

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

In this paper, a hybrid model called Xception Support Vector Machine (XPSVM) is proposed to predicts Short Term Traffic Congestion. Xception classifier mainly uses separable convolution, RELU and convolution techniques to predict feature detection in the dataset. Secondly, Support vector machine (SVM) classifiers use weighted regularization technique and fine-tuned binary hyperplane mechanism to predict output more accurately through maximum margin separation. Finally, the proposed model has been proved efficiency though the dataset which was taken from Google maps. The methodology is sound, however, the contribution of this study is not clear.

Major Concerns:

l  In the Introduction, there is a detailed description of the relative literatures in the paper. However, this section lacks a summary of the gaps of existing studies, resulting in the insufficient support of introduction for the contribution of this paper, which is logically abrupt.

l  In the model section, the author describes the specific form of the model, but does not elaborate on the applicability of the model. Why the model can be applied to traffic Congestion?

 

Some Minor Concerns:

l  The word “are” in “Traffic management are” should be replaced by “is”.

 

l  In the 5th line of 3th page of the paper, the word “a” in “a significant computation complexity” can be deleted.  

Reviewer 2 Report

 

The manuscript submitted to me for review is interesting from a methodological and empirical point of view. The research method and results do not raise any doubts. The authors clearly described how they obtained the data, which is an additional advantage. They took up quite an interesting topic which is congestion.

The conclusions were correctly formulated. The selection of literature is correct, although it could be extended.

The authors did not avoid minor shortcomings that need to be corrected.

The introduction should be revised to add to the purpose of the manuscript and there is no short description of the structure of the article.

The literature review needs to be organized. You need to somehow connect the sentences so that they form a coherent whole.

All variables should be described under the equations.

The manuscript editorial page needs improvement (various interlines).

The list of references must be adapted to the requirements of the publishing house.

 

Perhaps the authors will find additional inspiration in:

SZARUGA, E. (2020). Use of the artificial neural network to predict the energy intensity of means of public transport. In Rozważania o transporcie : księga jubileuszowa dedykowana profesor Elżbiecie Załodze, pp. 295-307. http://75lat.usz.edu.pl/wp-content/uploads/2021/04/ZALOGA_LOGO.pdf

 

Reviewer 3 Report

The research topic of this paper is Short Term Traffic Congestion Prediction. However, how to evaluate the traffic situation in a street or area need more criterions, not only considering in and out flow. More factors need be considered, such as parking, balance, etc. Therefore, the conclusion of this paper may not be correct.