Research On Analysis Of Collusion Behavior In Public Resources Trading Based On Data Mining

Public resources trading affects the national economy and people’s livelihood,and plays an important role in the rational allocation of public resources.However,in the course of the trading,the frequent occurrence of collusion behavior such as bid rigging affects the normal economic order of the public resources trading market seriously.Collusion behavior is extremely concealed,so it is difficult to supervise and manage it.In recent years,the integration of public resources trading types and the promotion of electronic bidding have accumulated massive trading data.Therefore,based on network theory,data mining technology is used to identify and analyze the collusion behavior in the process of public resources trading.Taking "collusion behavior in public resources trading" as the research theme,the purpose is to propose a set of methods that can scientifically and reasonably identify suspected collusion behavior in public resources trading.The specific research contents are as follows:Firstly,the research results related to collusion behavior in public resources trading are reviewed,and related concepts such as public resources trading and collusion behavior are defined,and the principles and application scenarios of social network theory,complex network theory,and data mining technology are clarified.Secondly,use web crawler to obtain public resources trading data,build a bidder relationship network,calculate and analyze the characteristic parameters of the relationship network.Thirdly,identify suspected collusion behavior in public resources trading.From the perspective of horizontal collusion,On the one hand,community detection algorithm and association analysis algorithm are used to screen bidders with obvious community structure in the bidder relationship network;on the other hand,function fitting analysis is used to screen bidders with obviously abnormal winning probability.From the perspective of vertical collusion,use correlation analysis to screen bidders and tenderers with strong correlation.Finally,set the level of supervision in the process of identifying suspected collusive behaviors,verify the effectiveness of the identification method,and put forward corresponding supervision suggestions.