**7. Discussion**

There is very little research on total risk assessment at the conceptual stage. For example, Oad et al. (2021) determined that price, the scope of work, and technical resources are the most important criteria in the bidding process at the conceptual stage [16]. Another study identified project cost, total project time, contract type, and use of risk management techniques as the primary criteria that can be used to assess overall risk in apartment buildings at the conceptual stage [15]. In this research, the main criteria used to estimate the total risks at the conceptual stage in residential projects are the number of floors, the building area, the interior finishes, the exterior finishes, the number of basements, the total duration of the project, the type of electromechanical, and the application of risk management processes. The scope of work that was identified as a critical factor in the study by Oad et al. (2021) was expressed in the current study by the number of floors, building area, internal and external finishes, and the number of basements. The duration of any activity can be estimated based on the required quantity and the production rate of the available resources. Hence, the technical resources identified by Oad et al. (2021) as critical factors in the conceptual stage were expressed in the quantities that can be inferred from the scope of work and the total duration of the project in the current study. Whereas the project cost component in the conceptual stage, which was identified by Badawy et al. (2022) can be estimated through many previous studies in the conceptual stage, which indicated that the cost can be deduced from the number of floors, building area, interior, and exterior finishes, and the number of basements, which was applied in the current study. Therefore, the eight input variables in the current study are in agreement with previous studies.

In the training phase, the proposed model predicted an average of 104 cases correctly and accurately with a ratio of 87.4% and predicted 15 cases incorrectly. The MAE for the low overall risk classification was 4.5%, and for the medium overall risk, the MAE was 14.9%. Unfortunately, the prediction of the overall risk in the case of the high-risk classification was 42.3%, which is considered a high ratio. In the testing phase, the proposed model predicted an average of 23 cases correctly and accurately with a ratio of 76.7% and predicted 7 cases incorrectly. The MAE for the low overall risk classification was 7.4%, and for the medium overall risk, the MAE was 14.3%. Unfortunately, the prediction of the overall risk in the case of the high-risk classification was 76.9%, which is considered a high ratio. Hence, the results indicated that this model is excellent in predicting the low and medium overall risk at the conceptual stage.

The mean absolute percentage error was 16.6% in an ANN model for estimating the total construction costs [26], while the MAPE was 13.2% in a hybrid technique for a cost assessment of residential projects at the early phase [21]. The MAPE was 26.3% an ANN approaches for cost forecast at the conceptual stage [32], while to estimate the cost of water treatment plants, the model has an error of 21.2% [29]. A model to predict the conceptual cost for engineering services in public construction projects was developed with a MAPE of 28.2 [30]. As a result of reviewing past research on conceptual-stage cost models, it was discovered that a mean absolute error of more than 13% was permitted, implying that the accepted model should have an error of less than 13%. The suggested strategy correctly classified 149 projects with a mean absolute error of 11.7%. Hence, this model can be accepted. The suggested model's acceptability implies that the eight input factors can be utilized to predict the overall risk of residential projects at a conceptual stage. The results of the study agreed with the viewpoint of the five experts who were interviewed to determine the most important criteria in the final list that can be used to predict the overall risk in the conceptual stage of residential construction projects. The most important of these factors was the number of floors, which represents 28.5%. The second top criterion was the interior finishes with 16.3 percent. The execution of the risk management process component ranked third, with 14.4 percent, while the floor area element came fourth, with 11.7 percent. The total project time was the fifth component that had a 10.8% impact on the overall risk forecasting in the conceptual stages, followed by the exterior finishes, which had a 10.2% impact. Finally, the electromechanical type had a weight of 6.2%, and the lowest parameter was the number of basements with a relevance of 1.7 percent. The importance of each component in determining the overall risk in the conceptual stages of residential projects is depicted in Figure 2.

**Figure 2.** Importance of parameters in estimating the overall risk.

### **8. Conclusions**

Decision-makers aim to predict the estimated value of the project budget in the conceptual stage to assess whether this investment is acceptable or not. The value of the reserve that covers the project's overall risk is included in the project budget. As a result, early on, a comprehensive risk assessment is required. There has been a great deal of research in developing cost estimation models, both in the planning phase and in the conceptual stages of a project. Unfortunately, there is a research vacuum in estimating risk in the conceptual stages of a project due to a lack of knowledge, so this study offers a model to forecast the overall risk at the conceptual stages of a project. A provisional list of essential characteristics, consisting of eleven parameters, was generated through a review of existing research and can be used to anticipate the overall risk in residential buildings at a conceptual stage. After three rounds of Delphi, the experts reached an agreement on the most critical parameters. The parameters for slab type, elevator number, and foundation type were omitted from the final list. Floor area, number of floors, interior finishes, external finishes, number of basements, kind of electromechanical, risk management process implementation, and overall project duration were all included in the final list. Four groups were created for

the floor space, the number of floors, interior finishes, and total project duration, while the internal finishes, the number of basements, the risk management method execution, and electromechanical kinds were all divided into two categories. Three levels were assigned to the output variable. Data were gathered from 149 actual residential projects. As a result, the confidence interval was 0.08 at the 95% confidence level. The model was simulated using artificial neural networks. The data were divided into five groups at random. There were twelve Multilayer Perceptron models identified and tested, each with a distinct number of hidden layers and activation functions. The proposed model has two hidden layers, the first of which has five neurons and the second of which has three neurons. In the suggested model, the Hyperbolic Tangent function was used to activate the hidden layer. The MAE was equal to 11.7% in this investigation. The number of floors is the most critical factor in determining the overall risk in the conceptual stages of residential projects, followed by interior finishing, and the risk management procedure. The electromechanical type and the number of basements were the least critical elements. The project manager can use the proposed model to identify residential projects in the conceptual stages as low-risk, medium-risk, or high-risk. As a result, the proposed model can assist stakeholders in deciding whether or not to continue with the project.

#### **9. Limitations of Research**

The overall risk and the influence of the important parameters were solely calculated based on the impact on the cost of the residential construction in this study. This study only looked at projects with fixed-price contracts. As a result, projects with cost-reimbursable contracts, for example, will require a re-estimation of the input parameter weights. The eight input criteria can be used in any country to obtain the overall risk at the conceptual phase. The data were obtained from 149 projects in Egypt, which means that the ranking of importance of each criterion may differ from one country to another. Hence, they should be double-checked the ranking of the importance of the criteria before being used in any other country. The user needs to alter the weights of the variables to adapt the model to subsequent times because the data used to produce it came from residential buildings in Egypt built between 2018 and 2020.

**Author Contributions:** Conceptualization, M.B. and F.K.A.; methodology, M.B.; software, M.B. and M.S.; validation, F.K.A.; formal analysis, M.S.; investigation, M.S.; resources, F.K.A.; data curation, M.B.; writing—original draft preparation, M.B.; writing—review and editing, F.K.A.; visualization, M.S.; supervision, M.B.; project administration, F.K.A.; funding acquisition, F.K.A. All authors have read and agreed to the published version of the manuscript."

**Funding:** This research was funded by the Research Supporting Project number (RSP-2021/264), King Saud University, Riyadh, Saudi Arabia.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data collected and used for analysis will be available from the corresponding author upon request.

**Acknowledgments:** The authors extend their appreciation to the Research Supporting Project number (RSP-2021/264), King Saud University, Riyadh, Saudi Arabia for funding this work.

**Conflicts of Interest:** The authors declare that there are no conflicts of interest regarding the publication of this paper.

#### **References**

