Font Size: a A A

Research On SVM-based Muti-bank Loan Pool Risk Analysis

Posted on:2011-03-26Degree:MasterType:Thesis
Country:ChinaCandidate:M YangFull Text:PDF
GTID:2249330374950079Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
The development of small and medium-sized enterprises (SME) is quite an important role for innovation of national economy, which make important contributions in solving the employment, technology innovation and stable financial. Mean the while, the development of SME are faced with many difficulties. Many studies and surveys show that hard to loan is important bottleneck for the development of SME. However, due to the capital market in our country at present is incomplete, banks in practice is still considered to be the most important channels of financing SME.Because its lower risk preference and the traditional mode of loan policy, banks actually often does not want to lend loans to SME. Hackethal and Gintschel put forward to combine different small banks’assets together, which constitutes many bank loans into the co-financing pool, participate gains and losses about banks under the multilateral agreement of co-financing pool.The key question of the method is the balance between risk of loan pool diversification effect and the participating bank’s effort decrease, the model not only encourage banks to participate in loan pool, but aslo encourge banks to pay the appropriate efforts to manage their own credit risk of loan assets in order to reduce the free-rider problem.This paper analysises the contract of the bank loan risk-sharing pool, obtains the closed-form solution, and discusses the features of the contract. The results show that multi-bank loan pool contracts can reduce small Banks’non-systematic risk caused by small size, reduce the information asymmetry between SME and banks, and provide a solution for SMEs in getting loans.This paper outlines differences of the risks faced by small and large enterprises by comparing differences in structure and management mechanism between them, and sums up the characteristics of risks which includes management risk, operational risk and financial risk. Then the paper selects11indicators of financial assessment and uses support vector machine as a tool to establish evaluation model of financial targets for SME which is considered to be a better solution due to solving small sample, nonlinear, high dimension and local minimum problems. This support vector machine overcomes defects of selecting parameters by experience by choosing cross-validation method and ideal penalty factor and the nuclear parameters. This paper compares support vector machine (SVM) model to BP neural network model which is currently thinking well precision warned. The result shows that using support vector machine to establish risk assessment model to predict is more accurate than neural network model.
Keywords/Search Tags:Small and medium enterprises, Risk characteristics, Financial early-warningindicators, Muti-bank Loan Pool, Support Vector Machine (SVM)
PDF Full Text Request
Related items