Research on Financial Risk Discrimination of Listed Companies Based on Fish School Algorithm

Because the financial risk data of domestic listed companies are unbalanced, and the traditional supervised model is difficult to establish a risk discrimination model through large sample learning. Due to the lack of data, the training of supervised models is usually based on manually filtered balanced data sets, which will affect the applicability of models in the real world. This research involves financial risk identification based on fish school algorithm. The this study is to compare the results of two algorithms (i.e. Fish School and Fuzzy Logic) to distinguish companies with different levels of financial risk. The comprehensively analyze the performance of two different methods in screening in PSX.