Risk Early-Warning for Enterprise´ Technological Innovation Based on Rough Sets and SVM

Technological innovation is a dynamic source that the enterprise acquires competitive advantage, and an important guarantee that enterprise tries for existence and development. However, it is a complex, dubious and high-risk activity, therefore, how to carry on early-warning risk of technological innovation has become a hot problem that enterprises study nowadays. In this paper, on the basic that new, scientific and logical risk early-warning index system of technological innovation is established, a hybrid intelligent system is applied to recognize the risk clusters of technological innovation, combining rough set approach and support vector machine (SVM). The reduced information table can be gotten with no information loss by rough set, and then, this reduced information is used to develop classification rules and train SVM to infer appropriate parameters. The rationale of the hybrid system is using rules developed by rough sets for an object that matches any of the rules and by SVM for one that does not match any of them. The effectiveness of our methodology was verified by experiments comparing BP neural networks with our approach.