Logo image
Using Backward-type Portfolio Selection Methods to Construct Optimal Portfolio Evaluated Index and Model
Thesis

Using Backward-type Portfolio Selection Methods to Construct Optimal Portfolio Evaluated Index and Model

Chien-Hsun Huang
Masters, 國立清華大學, 工業工程與工程管理學系
2003

Abstract

組合屬性 逆向式投資組合選擇 組合指標 多屬性混合整數二階規劃 多屬性二階規劃 Portfolio attributes Backward-type portfolio selection Portfolio index Multi-criteria MIQP model Multi-criteria QP model
Portfolio selection methods are developed in many fields. Many techniques and mathematical models are used to settle related problems based on mean-variance model developed in the stock markets. Many researches focus on evaluating items and formulate portfolio from good items and the methods belong to forward-type. On the contrary, this study aims to use “backward-type” portfolio selection method. In the perspective of backward-type selection, this thesis identifies the portfolio attributes into three categories such as independent, interrelated and synergistic portfolio attributes. Other than the mean-variance model considers the risk as the selected criteria. The thesis used the performance (i.e. future return) what the investor emphasized as the target. By the statistic of partial R squares from stepwise-regression method toward performance, the investors’ attitude (i.e. relative importance) of each attribute is obtained periodically and the evaluation index is constructed. Based on the index, the study then constructed multi-criteria mixed-integer quadratic programming model and quadratic programming by different definition of synergistic attributes to obtain invested position of stocks in the portfolio. Finally, This study will have illustrations in Taiwan Stock market and find that the backward-type selection methods, company profitability and synergistic attribute including in the model will have good performance.

Metrics

1 Record Views

Details

Logo image