The Computer Audit for Health Affordable Housing Based on Data Mining

Abstract

The study aims to improve the accuracy rate of data classification under computer audit. An appropriate data mining (DM) algorithm is employed to classify affordable housing data. First, the research status of the DM algorithm and affordable housing audits is introduced, and the process of computer audit is described. Second, after analyzing the advantages and disadvantages of well-known DM algorithms, a decision tree (DT) is constructed using C4.5. In the end, C4.5 is enhanced by the interest and simplified by the Taylor series. According to the findings, the naive Bayes tree (NBTree), the classification and regression tree (CART), and C4.5 have average accuracy rates of 85.4%, 85.7%, and 85.6%, respectively. This shows that C4.5 and NBTree have an excellent effect on classifying the sample set. CART and C4.5 have a faster modeling speed, and their classification accuracy on huge data sets is better than that of NBTree. Therefore, it can be concluded that C4.5 and CART have good scalability; C4.5 has the best performance, and its ratio of rules to leaf points is higher than that of the other two algorithms; compared to the original C4.5, the enhanced C4.5 has a greater accuracy rate and a rule-to-leaf node ratio, but their modeling time is the same; the simplified C4.5 has the best performance, its classification accuracy rate is 98.7%, its modeling time is 0, and its interpretability is 1.9. This study provides a reference for the information development of affordable housing audits.