A Study on Corporate financial crisis prediction models
| Autor: | Ming-Jen Lee, 李明仁 |
|---|---|
| Rok vydání: | 2008 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 96 After Beaver, Altman proposed corporate financial crisis prediction models, it was been study with human conern all the time, specially on parameter and method. This topic study on corporate financial crisis prediction models about forecast ability, stability, and usable value. We have three subjects in this research, （1）Using stock price target-EPS into discriminant analysis, logistic regression, cascaded logistic regression, back-propagation neural network, attempt to find out the models with better forecast ability.（2）Taking different proportion of samples into models to discuss forecast ability.（3）We discuss forecast ability, stability that building forward step logistic regression model, cascaded logistic regression models’ allow for category, cascaded logistic regression models’ without allow for category.We used holdout sample tests to show that the models’ usable value. This research finds as follow, （1）Stock price target-EPS factor has not effective to forecast ability in four models.（2）1：1 proportion of sample was better than 1：2 in type 1 error and models’ stability.（3）forward step logistic regression models bring a good forecast, but the usable value wrose than cascaded logistic regression models.In models’ forecast ability, type 1 error, type 2 error, usable value, cascaded logistic regression models’ without allow for category better than cascaded logistic regression models’ allow for category. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|