Applications of deep learning in stock market prediction: recent progress — part 5/5

Cleaned-text code points [62775, 65690). Bibliography omitted here; source.md includes it.

Most of the existing studies focus on only one stock market, in the sense that stock markets differ from each other because of the trading rules, while different markets may share some common phenomenon that can be leveraged for prediction by approaches such as transfer learning. There are already a few studies showing positive results for cross-market analysis [ Hoseinzade & Haratizadeh, 2019 , Lee et al., 2019 , Merello et al., 2019 , Nguyen & Yoon, 2019 , Hoseinzade et al., 2019 ] , it is worth exploring in the following studies. In Lee et al. [2019] , the model is trained only on US stock market data and tested on the stock market data of 31 different countries over 12 years. Even though the authors do not use the terminology of transfer learning, it is a practice of model transfer.

### 6.4 Algorithmic Trading

The prediction is not the end of the journey. Good prediction is one factor to make money in the stock market, but not the whole story. Some of the studies have evaluated the profit and risk of the trading strategies based on the prediction result, as we discussed in Section 4.4 . However, these strategies are simple and intuitive, which may be impractical limited by the trading rules. The transaction cost is often omitted or simplified, which makes the conclusion less persuasive. Another problem is the adaption for different market styles, as the training of deep learning models is time-consuming. These studies are not sufficient for building a practical algorithmic trading system. One possible direction is deep reinforcement learning, which has recent successes in a variety of applications and is also been used in a few studies for stock prediction and trading [ Xiong et al., 2018 , Lee et al., 2019 ] . It has advantages of simulating more possible cases and making a faster and better trading choice than human traders.

## 7 Conclusion

Inspired by the rapid development and increasing usage of deep learning models for stock market prediction, we give a review of recent progress by surveying more than 100 related published articles in the past three years. We cover each step from raw data collection and data processing to prediction model and model evaluation and present the research trend from 2017 to 2019. We also pay a special attention to the implementation of deep learning models and the reproducibility of published articles, with the hope to accelerate the process of adopting published models as baselines (maybe with new data input). With some future directions pointed up, the insight and summary in the survey would help to boost the future research in related topics.

## Acknowledgement

Weiwei Jiang: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Roles/Writing - original draft; Writing - review & editing.