Research on financial risk screening of listed companies based on clustering algorithm

The ability to recognize financial fraud activity in listed firms has grown in importance and is a constant source of worry in both academics and business. The A-share listed businesses in the biological and pharmaceutical industries that have faced penalties from the China Securities Regulatory Commission during the previous five years are used as samples in this study. Using Kangmei Pharmaceutical as an example, 24 characteristics are chosen using the fraud triangle theory, and testing and analysis are conducted using a Random Forest classification algorithm model in conjunction with SMOTE Oversampling technology. The findings show that it is more beneficial to use numerous feature sets to develop models or to construct models with many alternative algorithms for financial fraud screening studies, as opposed to merely classifying organizations into fraudulent and non-fraudulent categories.