PCA based pre-selection approach on portfolio selection problem

Abstract

This thesis investigates the portfolio selection problem. We proposed a pre-selection approach based on PCA to select stocks/portfolios before carrying out the online portfolio selection stage. In this work, we present several pre-selection methods that utilize the PCA method to compose new portfolios or select stocks as the new trading universe. We apply the PAMR as the algorithm in the online portfolio selection stage to conduct our experiments using the real trading data from Hong Kong stock market. The empirical analysis shows that we improvement the performance of the online portfolio selection regarding to the accumulated portfolio value and other measurements.

Date of Award	2021
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology
Supervisor	Jiheng ZHANG (Supervisor)