選股策略：結合技術分析與基本面分析

Abstract

本論文旨在證明透過結合不同面向的量化策略來建構投資組合是否可以有效提升其投資績效與穩定性。我們透過選用過去文獻所發掘出的不同領域之「異象」來組成混合策略投資組合，其篩選因子包含：過去一年累積報酬、獨特性風險、帳面市值比以及綜合性的獲利能力指標。透過結合以上技術分析與基本面分析的訊息，我們的實證結果顯示，混合策略在樣本期間能獲得顯著提升的超額報酬，且此結果與投資組合建構時篩選因子的順序無關。此外，混合策略也同時具有較高且相對穩定的夏普比率，其投資績效在樣本期間不但能打敗大盤，也顯著優於其它單一策略，此研究結果進一步顯示混合策略的有效性。

We use a variety of “anomalies” discovered in past literature to form the joint strategies to construct the investment portfolio. The screening factors include the past one-year cumulative returns skipping last month, the idiosyncratic volatility, the book-to-market ratio, and the composed profitability indicator. Our empirical result shows that, by combining the information of the above technical and fundamental analyses, we can obtain significantly improved excess returns during the sample period, regardless of the screening order of portfolio construction. Besides, the joint strategies also have outstanding and relatively stable Sharpe ratios. The investment performance of the joint strategies not only beats the market during the sample period but also significantly outperforms the other sole strategies. The main goal of this thesis is to show that combining different aspects of quantitative strategies can effectively improve its investment performance and stability.