Predicting long-term returns of individual stocks with online reviews

[摘要] Predicting long-term returns is essential for getting a full view of market efficiency. It is, additionally, more challenging for both human and algorithms, especially at the level of individual stocks, and competent solutions are still missing in previous effort. Considering the profound connections between stock prices and consumer opinions, asking experienced buyers can be a promising direction. In this paper, which is based exclusively on online reviews, a framework for predicting long-term returns of individual stocks is established. Specifically, 6,246 features of 13 categories inferred from more than 18 million product reviews are elaborated to build the prediction model. A satisfactory increase in accuracy, 13.94%, was achieved compared to the cutting-edge solution with 10 technical indicators being features. Additionally, we conduct a trading simulation on real-world stocks and our model obtains the highest return 13.12% among all baselines in half year. The robustness of the model is further evaluated and testified. Our approaches, in both feature inference and prediction model, will offer insightful advices to investors in long-term investments on individual stocks. (C) 2020 Elsevier B.V. All rights reserved.

[发布日期] 2020-12-05

[关键词] Long-term returns;Individual stocks;Online reviews;Predicting model;Trading simulation