ANN model predicting quality performance for building construction projects in Rwanda

Abstract

Quality is among the key factors of successful projects, which needs to be considered during the course of a construction project. Therefore, this study aims to assess the critical success elements contributing to quality performance and to develop an ANN model for predicting the quality performance within the building construction projects in Rwanda. By using a survey questionnaire, data collection about the application extent of 31 success factors identified by the mean of literature review and their corresponding quality performance were evaluated. Afterwards, SPSS and Python were used for the analysis. The significant factors affecting the quality performance of building construction projects in Rwanda were revealed and the model with best prediction was identified to be a feed forward neural network of one hidden layer and three hidden nodes based on back propagation algorithm with the prediction accuracy of 98.921% and the average cross entropy error of 0.016. This paper revealed the success factors affecting the quality and it can help to predict the quality performance for building construction projects in Rwanda and in the other countries with the same conditions for reducing risks that can results from poor quality performance.