Research on Stock Market Sentiment Analysis Methods and Models Based on Image Processing Algorithms in Financial Markets

View at publisher Summarise with VIRAT
Authors: Xing He
Cited by: 0
References: 37

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Aiming at the limitations of traditional technical analysis methods in stock market sentiment analysis that rely on subjective judgment and manual feature design, this study proposes an end-to-end sentiment classification method based on deep convolutional neural network. The method converts stock OHLCV trading data into standardized K-line images, integrates three levels of visual information: shallow geometric details, mid-level morphological patterns and deep semantic abstraction through a multi-scale feature fusion mechanism, and uses a weighted cross-entropy loss function to deal with the class imbalance problem to realize the automatic recognition of bullish, bearish and neutral emotions. Experimental results verify the significant performance advantages of the proposed method over traditional machine learning and the existing deep learning baseline methods. Confusion matrix analysis and feature space visualization further confirm the good discriminative ability of the feature representation learned by the model. The research shows that the fusion of image processing algorithms and deep learning technology provides an effective solution for financial time series sentiment analysis, and the method has practical application value in the development of investment decision support systems and quantitative trading strategies.

Fields of Study & Subjects

Stock Market Forecasting Methods Sentiment Analysis and Opinion Mining Financial Markets and Investment Strategies

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
1548–1558
Publication date
28 January 2026
Language
EN