The successive opening and liberalization of the financial market has led to the rapid development of the capital market, which has led to the occurrence of financial crises. The research data comes from Taiwan Economic Journal (TEJ), the sampling period is from 2000 to 2021. This study uses a two-stage way to construct the model. First, the decision tree C5.0 and random forest are used to screen out important variables, and then the convolutional neural network (CNN) and long short-term memory (LSTM) model are used to construct an effective financial crisis prediction model. This study will make effective suggestions for academic research and practice.
