Books on Neural Networks and Deep Learning for Stock Forecasting
Summary
The discussion responds to a request for learning materials on stock-price forecasting with artificial neural networks, LSTMs, regression, and other predictive methods. It recommends two practical books: one covering modern forecasting models, including LSTMs and neural forecasting systems, and another focused on deep learning for time series. The respondent values their code examples as a way to try implementations on data.
These are recommendations rather than a comparative review or evidence that the methods produce profitable forecasts. The author says their return to the literature is recent, so the endorsements are personal and tentative. The answer also expresses skepticism about dependable price predictability and points to competition from existing algorithms, particularly faster high-frequency systems. It offers no tested forecasting workflow, performance results, or guidance on data preparation, validation, transaction costs, or avoiding overfitting, so readers would need other sources to assess whether a model generalizes to trading.
Key ideas
- Practical books on time-series forecasting can introduce LSTMs and other neural forecasting methods.
- Code examples make it easier to experiment with implementations on financial data.
- The recommendations are personal and are not supported by a comparative evaluation.
- Forecasting accuracy does not by itself establish that a strategy can earn profits after competition and costs.
Tags
Full text
# Book/Material recommendation - Stock Price Forecasting using AI tools # Book/Material recommendation - Stock Price Forecasting using AI tools I am looking for a book, which covers the following topics: - stock price prediction using Artificial Neural Network, - stock price prediction using LSTM, - stock price prediction using linear/non-linear regression methods, - other prediction techniques. I am not interested in Black-Scholes, or other quant stuff. I need book/materials on stock price predictions. Thanks! ## Answer by Konstantinos (score 1, accepted) https://quant.stackexchange.com/a/66014 After some decades of absense from the meaningless price chasing (money don't bring happiness), the recent squeeze seasons lured me into exploring some of the literature. Hence, my opinion is not robust. Anyway, so far my favorites are: Korstanje, 2021, "Advanced Forecasting with Python: With State-of-the-Art-Models Including LSTMs, Facebook’s Prophet, and Amazon’s DeepAR" and Brownlee, 2018, "Deep Learning for Time Series Forecasting" because they include the code for some fast implementations and I can easily play with the data. They seem to cover everything you asked. If I do not lose interest I would like to study some more "science-y" material. The top of my list is Velu et al., 2020, Algorithmic Trading and Quantitative Strategies (which is unrelated to your question, I just mention it). All in all, I don't think there is predictability, perhaps reactability. There is an enormous web of algorithms already in use and you cannot do anything, and even in the 1 in the million+ that you can, a high-frequency algorithm will be faster and ahead of you. :D
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.