Prediction of construction projects' costs based on fusion method. Issue 7 (2nd October 2017)

• Record Type: Journal Article
• Authors: Shutian, Fang; Tianyi, Zhao; Ying, Zhang
• Is Part Of: Engineering computations. Volume 34:Issue 7(2017)
• Page Start: 2396
• Page End: 2408
• Publication Date: 2017-10-02
• Subjects: ANN -- Kalman filter -- Multiple linear regression -- Construction project's cost prediction -- Least square-support vector machine; Computer-aided engineering -- Periodicals; Computer graphics -- Periodicals; 620.00285
• DOI: 10.1108/EC-02-2017-0065
• Languages: English
• ISSNs: 0264-4401

Abstract

Abstract : Purpose: This study aims to predict the construction cost in China, the authors purposed a fused method. Design/methodology/approach: The authors extracted 22 factors which may influence the cost and performed the correlation analysis with cost. They chose the highest 10 factors to predict cost by the fused method. The method fused the Kalman filter with least squares support vector machine and multiple linear regression. Findings: Ten factors which affect the cost most were found. The construction cost in China can be predicted by the presented method precisely. The statistical filter method could be used in the field of construction cost prediction. Research limitations/implications: The construction cost and construction interior factors are a business secret in China. So, the authors only collected 24 buildings' data to perform the experiments. Practical implications: There is no standard and precise method to predict construction cost in China, so the presented method offers a new way to judge the feasibility of projects and select design schemes of construction. Originality/value: The authors purposed a new fused method to predict construction cost. It is the first time that the statistical filtering method was used in this field. The effectiveness was verified by the experiments. Ten factors which have a high relationship with construction cost were found.