Research On Classification Rules Of Affordable Housing Audit Based On Decision Tree Algorithm

In recent years,with the development of information technology,the trend of the data of the audited units is very obvious.This puts forward new requirements for the audit work,how to find more valuable audit information from the vast amount of computer data has become a new problem.In this case,the data mining technology was experimented with the introduction of the computer audit,combination of technology and expertise of the audit auditors can find something beyond our experience of knowledge through data mining,can significantly improve the audit efficiency and accuracy.According to the audit bureau of ZhuMaDian city in 2016,low-income housing project audit,illegal mining for low-income housing to personnel from the previous algorithm found the relevant information for learning in violation of the general rules of low-income housing for data mining classification rules of the illegal application.Due to the obvious classification of mining tasks,this paper focuses on the study and comparison of the typical classification algorithms and their application in audit.According to the characteristics of continuous data mining and the advantages of decision tree model,after comparing the performance of several common decision tree algorithms,the C4.5 algorithm is used to extract the classification rules from a large number of affordable housing data.Then,this paper will be collected into the relevant departments of the data into the database,which includes the city's vehicle information,pension insurance information,provident fund loan information,business registration information,public rental information.But these data have different degrees of incomplete,noise and inconsistencies,followed by data cleaning,transformation,integration of pre processing technology for the processing of the original data,and finally select attribute mining related theme that unified view.Given the working principle and method of evaluation of C4.5 algorithm classification result is described in detail,this paper carries out the analysis of the data after pretreatment of the low-income housing using the C4.5 algorithm,constructed a decision tree model,and then extract the classification rules are easy to understand,and the accuracy of the classification results of the test method to keep.The experimental results prove the applicability of the C4.5 algorithm in the classification of the rules for the protection of the housing violations.Finally,according to the C4.5 algorithm in attribute selection deficiencies,this paper introduces the concept of degree of interest,on the C4.5 program made the change of the selected attribute information gain is modified,the relative change in support of decision attribute.The improved C4.5 algorithm is more suitable for the auditors to build the decision tree model of the same training set,and its accuracy is also improved,which proves the effectiveness of the improved algorithm.However,through the analysis of the improved C4.5 algorithm is C-C4.5 algorithm,although this algorithm is improved in accuracy,but the complexity of the cost is high,therefore,the C-C4.5 algorithm uses the principle of Taylor series and the equivalent infinitesimal was further improved,the formation of the C2-4.5 algorithm,and the validity is proved by the improved contrast.In this paper,through the extraction of the illegal application of classification rules in low-income housing,low-income housing in 2016 ZhuMaDian city audit effectively improves the efficiency of audit subjects in the field investigation,to improve the accuracy of the audit doubts to explore,discover the false information reporting(such as real estate information fake),illegal for low-income housing 456 sets,and results He Nan provincial audit department in 2016 outstanding computer audit case.