An enterprise crisis predicting system based on outlier data mining

Abstract

Many factors in an enterprise are playing important roles in intense commercial competitions. Some are positive and some are negative. If dealing with these factors correctly, potential crisis will be found to avoid defeats, even bankrupt. Designing a crisis predicting system is necessary. An excellent predicting can not only predict expecting crisis and take controlling measures, but also can provide enough preparation and plan to deal with crisis smoothly. The factors are the basis data to be analyzed to support such a system and maybe they are quantitative or qualitative. In order to solve such problems as half-structured and non-structured data analysis in enterprise crisis predicting system, a predicting system based on outlier data mining is put forward. The system organization, frame construction, function and working principles are illustrated. And the working process is showed by an example of cheat predicting. The experimental results show that this method is efficient and it has wide utilization in predicting fields.