3.1.1. Gathering the MLP Network's Data

Data regarding road construction projects in Fars Province from 2010 to 2020 were gathered by conducting a vast study on documents from Fars Road Administration. Then, data were turned into matrices. In order to determine the best pattern of network input, all the probable factors affecting the cost of projects were determined. Thus, as input, 14 neurons were formed. These neurons were the identified factors, namely, "Payment status (F1)", "Climate (F2)", "Conflicts (F3)", "Plans (F4)", "Accessibility of materials and appliances (F5)", "Fortuitous events (F6)", "Delivery of land (F7)", "Minor contractors (F8)", "Project schedule (F9)", "Qualification of project management team (F10)", "Inflation rate (F11)", "Risk management (F12)", "Relationship among project's parties (F13)" and "Initial geotechnical studies (F14)". Moreover, the output layer of the ANN was determined, i.e., the cost performance index of the projects.
