|
BUILDING COST INFORMATICS WEB-BASED NEURAL NETWORK-REGRESSION HEURISTICS PROTOCOL FOR DECISIONS IN BUILDING CONSTRUCTION PROJECTS |
|---|---|
| รหัสดีโอไอ | |
| Title | BUILDING COST INFORMATICS WEB-BASED NEURAL NETWORK-REGRESSION HEURISTICS PROTOCOL FOR DECISIONS IN BUILDING CONSTRUCTION PROJECTS |
| Creator | Lekan M Amusan, Ignatius O Omuh, Timothy O Mosaku |
| Contributor | Didem Ozevin, Hossein Ataei, Mehdi Modares, Asli Pelin Gurgun, Siamak Yazdani, Amarjit Singh |
| Publisher | ISEC PRESS |
| Publication Year | 2562 |
| Keyword | Coefficient, Informatics, Collinearity, Fitness, ICT, Updating, Expert System |
| Abstract | 1. Building Informatics is a body of knowledge that uses ICT computer system, digital system, building information modeling and state of art software in solving technical and management issues in building and construction fields. The data was generated through the combination of parametric regression method and neural network (an expert system). One of the modern methods used in data forecasting and modeling is Artificial Neural Network considering its advantage over traditional regression method. Data sample of One thousand and five hundred (1500) residential building projects completion costs, out of original data trained with an Artificial Neural network were selected at random and divided into two half as sample for the study 2. one part was used to generate the network heuristic algorithm while the second part was used in modeling and simulation of the system for model validation. Regression analysis was carried out and model validated with functionality and Jackknife re-sampling technique. 150 Questionnaires was used to capture data on factors influencing application of heuristics protocol for decisions in residential building construction projects and data samples were analyzed using severity index, ranking and simple percentages. Analysis of data brought to the fore some factors that influences effective application of heuristic protocol in solving decision problems in construction decision process. The linearity analysis was carried out on the model and results indicate high level of tolerance and - 0.0876 lowest variation prediction quotients to 0.9878 highest variation quotients. Also 0.069 regression model fitness coefficient (R-square) was generated with 0.9878 highest variation quotients with standard error of 0.045. The results data attest to the stability of the model generated and the model is flexible in accommodating new data and variables, thus, allows for continuous updating |
| ISBN | 978-0-9960437-6-2 |
| Language | English |
| URL Website | https://www.isec-society.org/ |
| Website title | ISEC Society |
