Reasons and control strategies of construction project over budget based on big data analysis

Abstract

During the construction process of construction projects, the cost budgeting work is an important guarantee for the entire project construction. Due to the influence of the social and economic market and the error problems caused by the related cost staff during the work process, these external factors will affect the construction project. The overall cost budgeting work has had a serious impact. The internal cost of construction projects has received the full attention of the company's internal management staff. It is mainly aimed at common problems and loopholes in cost work, which is manifested as the phenomenon of over budget in engineering. In this case, serious problems such as out-of-control costs in building construction will also arise, and eventually the overall economic benefits of construction enterprises will decline. Based on the above background, the purpose of this paper is to study the reasons for the over-budgeting of construction costs and their control strategies based on big data analysis. This article mainly addresses the problem that traditional anomaly data detection methods cannot detect anomalous data that is related to the description of the list. This paper proposes to use a list classifier and use the clustering center obtained by K-means clustering as a label. This kind of abnormal data is detected, and the experimental results of the proposed method are compared with traditional methods. Among them, the accuracy of the method of using the list classification method to detect abnormal data of the integrated unit price is 0.9107, while the accuracy of the traditional distance-based method is 0.8304. The experimental results show that the method proposed in this paper using the inventory classification method to detect abnormal data of the integrated unit price is more accurate than the traditional method, and can solve the problem of over-budgeting of construction costs to a certain extent.