| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a bankruptcy prediction system which uses quantitative data. <P>SOLUTION: Prediction data are provided at least through a step (a) of gathering enterprise information, evaluating individual enterprises, and storing evaluation data in a database in response to input from an operation terminal, a step (b) of selecting data suitable for model construction from the evaluation data through operation from the operation terminal and converting the data into an item fuitable for data mining, a step (c) of taking logistic regression analysis using the converted data and selecting a variable persuasive to 'bankruptcy' and a method for variable processing, a step (d) of generating a plurality of kinds of models whose internal adopted data are correspondingly persuasive, applying prediction results of the respective models to the individual enterprises for which an object period has passed, and measuring the operation precision of the models, and a step (e) of gathering the latest information on the individual enterprises and reviewing bankruptcy predicted values every time when evaluation is reviewed or negative information is gathered. <P>COPYRIGHT: (C)2003,JPO</p> |
