Machine learning for predicting construction activities schedule estimation and actual completion in Myanmar | |
| Author | Sai Sai Hseng Khoe |
| Call Number | AIT Thesis no.CM-23-14 |
| Subject(s) | Machine learning--Burma Construction projects--Burma--data processing MATLAB |
| Note | A thesis submitted in partial fulfilment of the requirements for the degree of Master of Engineering in Construction, Engineering and Infrastructure Management |
| Publisher | Asian Institute of Technology |
| Abstract | This research is focused on identifying the important factors that affect the activities duration and to predict the estimated duration and actual duration of the activities by developing the Artificial Neural Network model in MATLAB software. The study is focused on Myanmar. The data was collected form medium size building construction projects in Myanmar. The study was conducted by interviewing 5 experts who are working in Myanmar construction projects. The factors that affect the duration of activities were listed based on the literature review of previous research and expert interviews. Then 5 experts were given to fill the expert validation based on each expert. For the factors that affect the activities duration as one of my objectives is satisfied. A neural network model has been developed in MATLAB. A configuration of 21-10-1 neural network was made, which has 21 input parameters and 1 output parameter for estimate duration of earthwork excavation. The model is trained with 10 hidden neurons with Sigmoid as its activation function. After training the model, it showed an accuracy of 85.09%. A configuration of 22-10-1 neural network was made, which has 22 input parameters and 1 output parameter for estimate duration of concreting the foundation. The model is trained with 10 hidden neurons with Sigmoid as its activation function. After training the model, it showed an accuracy of 93.64%. A configuration of 14-10- 1 neural network was made, which has 14 input parameters and 1 output parameter for predicting the actual duration of earthwork excavation. The model is trained with 10 hidden neurons with Sigmoid as its activation function. After training the model, it showed an accuracy of 94.47%. A configuration of 14-10-1 neural network was made, which has 14 input parameters and 1 output parameter for predicting the actual duration of concreting the foundation. The model is trained with 10 hidden neurons with Sigmoid as its activation function. After training the model, it showed an accuracy of 83.31%. |
| Year | 2023 |
| Type | Thesis |
| School | School of Engineering and Technology |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Construction Engineering and Infrastructure Management (CM) |
| Chairperson(s) | Hadikusumo, Bonaventura H. W. |
| Examination Committee(s) | Kunnawee Kanitpong;Wasan Teerajetgul |
| Degree | Thesis (M. Eng.) - Asian Institute of Technology, 2023 |