Provided is an enterprise default prediction apparatus using synthetic feature generation and a method thereof, wherein the enterprise default prediction apparatus includes a memory in which the enterprise default prediction program is stored and a processor executing a program stored in the memory, and the processor is configured to execute the enterprise default prediction program. Accordingly, a plurality of basic features are extracted based on a plurality of financial indices representing the financial conditions of the companies, and two basic features are randomly selected from the extracted plurality of basic features according to a randomly selected arithmetic operation. Calculate the composite features by calculating, calculate the feature importance, which indicates the degree of impact of each feature on the company's default, based on the basic features, composite features, and financial information representing the financial condition of the default companies, and calculate the Basic or synthetic features with high importance To be pre-selected as the set number to perform a selection of default features or enterprise bankruptcy prediction based on the synthesized characteristic.
展开▼
