A Classification of Mutual Fund Customers Risk Assessment Based on Gene Expression Programming

• Main Authors: Lu,Hui-ching, 呂彗青
• Other Authors: 林文修
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/94166080363746521289

碩士 === 輔仁大學 === 資訊管理學系 === 98 === Mutual fund is one of popular investments and financial company should improve customer protection through investor classification based on risk assessment with information technology. This process benefits both financial companies and their customers. The research objective is to study if GEP (Gene Expression Programming) can apply on investor classification. The variables of this research are sex, age, education, career, living location, experience on investment, investment amount, risk assessment, and mutual fund type. The research shows rate of accuracy up to 97.73% with GEP to predict customers risk assessment. GEP(Gene Expression Programming) has good learning capability and obtain the best solution from different combinations. In summary, accuracy ups to 97.728, SE ups to 98.795%, and SP is 94.780 with GEP and GEP performs better in SE than other two models. The research shows that is it better to use GEP to predict customers risk assessment.