Using machine learning method to improve the performance of smart beta Portfolio

• Main Authors: Cheng,Jen Chang, 鄭人彰
• Other Authors: Huang,Hsian-Chang
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/n582u9

碩士 === 國立彰化師範大學 === 企業管理學系 國際企業經營管理(IMBA) === 105 === Active and passive portfolio fund management and performance has been compared and discussed for decades. Many studies reveal passive portfolio funds has better performance then active portfolio funds and the transparency of active portfolio funds stock selection process also been doubted by investor. So the passive index portfolio funds become popular in stock market in recent years and the market capitalization is rising. For tracing the market index purpose , most of the passive index funds invest in specific large cap company stocks which will highly affect the performance of passive index portfolio funds. For improve the performance of passive portfolio index fund, alternative strategy for passive portfolio index fund is develop in funds market, we called Smart Beta fund. The Smart Beta funds invest in specific factors and combine active and passive investment strategies. For improve the performance of Smart Beta, we use machine learning method to study the best strategy for Smart Beta funds. The results show that LIBSVM can improve the performance of the Smart Beta funds when we apply specific financial data. This means machine learning method will assist fund manager to build a better portfolio in stock market.