DATA QUALITY ISSUES THAT HINDER THE IMPLEMENTATION OF ARTIFICIAL NEURAL NETWORK (ANN) FOR COST ESTIMATION OF CONSTRUCTION PROJECTS IN MALAYSIA

Abstract

The Artificial Neural Network (ANN) which is one of the Artificial Intelligence (AI) tools, has been identified as a great technique to be used for construction cost estimation in the project. With the optimum quality of data input into the ANN model, it could produce an optimum and reliable cost estimation output. Nonetheless, the construction industry is lack of breadth and depth of data that is required as input into ANN. Though many online databases have been made available for data consumers, data quality problems remain unresolved. Thus, this study aims to identify data quality issues that can hinder the implementation of ANN for cost estimation of a construction project. A literature review and semi-structured interview were employed for the data collection of this research. The content analysis method was used to analyse the information obtained through the literature review. Meanwhile, the data collected from the semi-structured interview with nine (9) respondents were analysed using both content analysis and descriptive statistics analysis methods. The findings revealed six data quality issues that can hinder the ANN implementation for cost estimation of construction projects in Malaysia which are inaccurate data, outdated data, data access barriers, insufficient data, noise in training data, and data input degree of influence. Academically, this study contributes to the body of knowledge about the implementation of ANN for cost estimation of construction projects in Malaysia.

Keywords: Artificial Neural Network, Artificial Intelligence, Cost Estimation, Data Quality.