| Summary: | 碩士 === 輔仁大學 === 統計資訊學系應用統計碩士在職專班 === 105 === Enterprises and investors are concerned with financial crisis, but investors can only learn the financial situation from companies’ financial report. Due to this reason, the construction of warning model is an important issue in receut years. The purpose of this study is to construct the early warning model of financial crisis for traditional industries, so that the risk management departments and investors can refer to the models for their investment and know the risk in the early stage. In this studey, two types of warning models are considered, the one-stage and two-stage models. The results show that, regardless of financial variables or non-financial variables, the companies having solid finance show better performance than the companies having financial crisis. In addition, the research selects 11 significant variables which affect the probability of financial crisis. It is found that the one-stage neural network models has the best accuracy.
|
|---|
