Expert System-Based Predictive Cost Model for Building Works: Neural Network Approach
L. M. Amusan, T. O. Mosaku, Charles Korede Ayo, Albert B. Adeboye · Covenant University Repository (Covenant University) · 2013
Project managers need accurate estimate of building projects to be able to choose appropriate alternatives for their constructions. Estimated costs of building projects, which hitherto have been based on regression models, are usually left with gaps for high margin of errors and as well, they lack the capacity to accommodate certain intervening variables as construction works progress. Data of past construction projects of the past 2 years were adjusted and used for the study. This model is developed and tested as a predictive cost model for building projects based on Multilayer Perceptron Artificial Neural Networks (ANNs) with Levenberg Marqua. This model is capable of helping professionals save time, make more realistic decisions, and help avoid underestimating and overestimating of project costs. The model is a step ahead of Regression models.