Credit Scoring System Design Based On Support Vector Machine

With the development of economy improved rapidly in our country recently,Peer-to-peer lending websites are at the leading edge of a microcredit boom in China.Credit risk management is one of the most important issues in financial research. Reliablecredit scoring models are crucial for financial agencies to evaluate credit applications.Credit evaluation model have been widely studied in the field of machine learning andstatistics.Support Vector Machine (SVM), which is mainly used in pattern recognitionproblems at the beginning, is a fresh machine study method put forward by Vapnik usingstatistics principles in early1990s. SVM is a new method in data mining. It has excellenttheory foundations, which are structure risk minimization, conditional optimization theoryand kernel space theory. In order to solve a complicated classification task, the core idea isthat SVM maps the vectors from input space to feature space in which a linear separatinghyper-plane is structured.The thesis is four parts mainly mark, part I firstly introduced the relevance conceptand the academic research about credit scoring; Part II described the definition of SVMand F-score, we also proposed a new credit scoring model based on feature weighted SVM,the experiment demonstrated the effectiveness of the classification algorithm; Part IIIdiscussed about the structure of the system; The last part reviewed the research work andpointed out the direction for future research|.