Abstract We consider stock price charts as images and use deep learning neural networks (DLNNs) for image modeling. DLNNs can imitate the work of a technical analyst to predict stock price movements in the short term with price charts and stock fundamentals (e.g., price-to-earnings ratio). We find that a deep learning model performs better than a single-layer model in the prediction of the Chinese stock market. DLNNs provide customizable statistical tools for analyzing price charts effectively. More importantly, price trends established by different periods of past daily closing prices dominate stock fundamentals in predicting future price movements.
Paper
Full text
Stock market prediction with deep learning: The case of China
Semantic Scholar · Computer Science · 2021
Abstract
Abstract We consider stock price charts as images and use deep learning neural networks (DLNNs) for image modeling. DLNNs can imitate the work of a technical analyst to predict stock price movements in the short term with price charts and stock fundamentals (e.g., price-to-earnings ratio). We find that a deep learning model performs better than a single-layer model in the prediction of the Chinese stock market. DLNNs provide customizable statistical tools for analyzing price charts effectively. More importantly, price trends established by different periods of past daily closing prices dominate stock fundamentals in predicting future price movements.