AN ANALYSIS FOR PREDICTION OF DELAY OCCURRENCES IN BUILDING MANAGEMENT SYSTEMS IN SRI LANKA

Statistics

Predicting whether a delay arises in a BMS project, assists with early resource and time allocation. This study intends to provide BMS project managers and clients at an early stage of a project, an accurate yes/no on whether a delay could occur. Through descriptive statistics, variables with significance to delay occurrence are identified. The main findings from the descriptive statistics were; out of all quantitative variables, the number of points and the estimated time of a BMS project were identified as significant by dot plot diagram. From the Chi-square test for association done for all categorical variables, client type proved to be significant with a p-value of 0.037 (<0.05) for the association between the variables, client type and delay occurrence. Considering these significant variables, linear discriminant models and binary logistic models are constructed. Both types of models will be validated by k-fold validation and the most accurate model will be chosen using a classification table. The binary logistic models considered all significant variables but, the linear discriminant model was constructed by only the quantitative predictors as per the assumption of constructing such a model. Therefore, only the number of points and the estimated time of a BMS project was considered to build the linear discriminant model. The most accurate model was identified as a linear discriminant model constructed in the 3rd fold of delay occurrence vs. points and the estimated time had the best accuracy (0.8571). This model proved more accurate than all models constructed in the study. The accuracy reveals that a linear discriminant function is more accurate to predict the occurrence of a delay in BMS projects than a binary logistic model.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.