A User Profile And Monitoring Early Warning Of Tendering And Bidding Corruption In The Construction

Since the 14th Five Year Plan proposed that China will organize and implement a batch of major projects related to infrastructure,people’s livelihood construction,and other fields,engineering construction has ushered in a new round of vigorous development.However,during the period of great development and construction,corruption issues are also prone to high incidence,which is not uncommon in construction project bidding,seriously damaging fair market competition,and posing a major threat to the sustainable development of China’s construction industry.In recent years,government departments have carried out various digital special rectification activities against bidding corruption.However,as bidding corruption becomes increasingly diverse and covert,targeted efforts and precise measures for bidding corruption governance have become difficult.It is urgent to clarify the laws of bidding corruption and monitor the corrupt subjects and behaviors in a targeted manner.In order to improve the utilization rate of corruption data and explore the hidden corruption information behind the data,this article is based on the case data of bidding documents for the construction project of China Judgment Document Network.939 bidding corruption judgment documents and 1014 corrupt behavior person information were collected from China Judgment Document Network.Based on literature review and actual case characteristics,a label indicator system was constructed,The user portrait technology is used to depict the portrait characteristics of the bidding corruption actors on different dimension labels.In order to grasp the specific behavior of corruption,the LDA topic model is used to obtain six main types of corruption behavior and specific behavior The main types of behavior are bribery and help,fraud and fraud,collusion bribery and profiteering,and abuse of power.The differences in the subject attribute characteristics of bidding corruption are summarized.The unstructured problem of case data has been effectively addressed through user profiling analysis.Further,referring to the commonly used recommendation algorithms in the research of recommendation systems,a collaborative filtering(CF)recommendation model based on similarity calculation of user basic attributes and behavior is constructed to predict the behavior of bidding corruption subjects.The effectiveness of the model is verified through case analysis.Based on the case library,a large screen for monitoring and warning corruption in bidding has been designed to empower the implementation of digital anti-corruption.From the perspective of user profiling,multi-dimensional disclosure and characterization of the subject and behavioral characteristics of bidding corruption can be made,predicting the tendency of the target subject’s corrupt behavior,and promoting the transformation of corruption governance from post attack to pre monitoring and prevention.Based on the overall characteristics of bidding corruption,targeted monitoring and carly warning suggestions have been proposed from the aspects of monitoring systems,subject supervision,and policies and regulations,to assist in the transformation of traditional bidding corruption governance to the use of big data precise supervision,and to take multiple measures to ensure the efficient and clean promotion of the "14th Five Year Plan" development and construction.