The Study Of Period Gene Model Method Based On The Early-warning Of Financial Distress

The theory and methods of corporate financial distress prediction have originated in the 1930s, and there has been many mature theory, methods and tools to explain and predict corporate financial distress via many years of scholars’research. There are more representative of the quantitative research methods: single variable analysis, multivariate discriminant model, Logistic regression model, neural network models, genetic algorithms, support vector machines, decision trees, and so on. Most of these classic models first select the indexes under a certain method, which have smaller relation and are in line with certain principles. Then these indexes are put in the pointer type classifier (computational model) to classify and determine the sample. These methods are common: (1) indexes entered the calculation model less; (2) lack of analysis for the trends of financial indicators over time which could be a research object. The method of period gene model takes two points into account in this paper.Based on the results of previous studies, this paper proposed the concept of the period gene model via introducing the model theory in genetic algorithm and combining with the concept of time series, which converted the changing trend of financial indexes to the period gene model in a certain data conversion rules. More financial indexes were selected, which were conducted large-scale time-series analysis used statistical methods. At last, several characteristic time gene sub-models were gained. The article designed the period genetic map, which expressed the period gene model’s information. The several representative period genetic models were found, which had high prediction accuracy. It demonstrated that the method of period genetic model was fit for the time-series analysis of financial indexes and it had a good effect.