Application of Tax Risk Management System based on Hybrid Association Rule with Apriori Algorithm

In China, listed manufacturing businesses are one of the major sources of tax income. Similarly, the complexity of tax collection and administration is rising along with the diversity of taxes and the rules and regulations. So, building a methodical and efficient strategy for identifying and evaluating tax risk become crucial. This research proposes a hybrid Association Rule with Apriori algorithm (ASR-AP) to construct the audit data mining system to evaluate tax compliance accuracy under the presumption of difficult tax work. After processing the crawling financial index data from listed manufacturing businesses, the analysis of corporate tax compliance are performed using Apriori algorithm. After splitting the risk levels of tax compliance across the businesses, it is determined that weak profitability, a poor capacity to pay off debt, and a high percentage of past-due repayments are shared by businesses with low tax compliance. When compared to traditional models such as Support Vector Machine (SVM), Hybrid Unsupervised Outlier Detection (HUNOD) Autoencoder, the proposed ASR-AP achieved higher accuracy of 99.40%.