Stock Prediction Based on Principal Component Analysis Principal Component Analysis and Long Term Short Term Memory

Wenxuan Liu

2023

Abstract

Nowadays, machine learning mode in financial markets has become popular. This experiment studied the potential benefits of integrating Principal Component Analysis (PCA) with Long Short-Term Memory (LSTM) neural networks for the prediction of stock prices. The data used in the experiment are collected from Yahoo Finance, a reputable platform for collecting stock prices. Traditional methods for stock price prediction by using LSTMs mostly just rely on the raw historical stock data. Raw data are high-dimensional and may contain redundant information. The redundancy could reduce the model’s predictive ability. The hypothesis claims that the application of PCA can refine the data and enhance the predictive performance of LSTMs. To prove the hypothesis, I employed three steps: Initially, I applied PCA on the historical stock data to preprocess the principal components. In addition, use these components as inputs for the LSTM model. Lastly, compare the performance of the PCA-integrated LSTM model with a traditional LSTM model that uses unprocessed data. In the result, compared with the original stock data, the prediction accuracy of the LSTM model trained using PCA-converted data has been significantly improved. The result not only can prove the hypothesis but also underscores the advantages of combining dimensionality reduction techniques with the LSTM method.

Paper Citation

in Harvard Style

Liu W. (2023). Stock Prediction Based on Principal Component Analysis Principal Component Analysis and Long Term Short Term Memory. In Proceedings of the 1st International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-705-4, SciTePress, pages 392-396. DOI: 10.5220/0012816500003885

in Bibtex Style

@conference{daml23,
author={Wenxuan Liu},
title={Stock Prediction Based on Principal Component Analysis Principal Component Analysis and Long Term Short Term Memory},
booktitle={Proceedings of the 1st International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2023},
pages={392-396},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012816500003885},
isbn={978-989-758-705-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Stock Prediction Based on Principal Component Analysis Principal Component Analysis and Long Term Short Term Memory
SN - 978-989-758-705-4
AU - Liu W.
PY - 2023
SP - 392
EP - 396
DO - 10.5220/0012816500003885
PB - SciTePress