A Multimodal Aspect-Level Sentiment Analysis Method Based on Syntactic Enhancement and Aspect-Oriented Approach

By constructing a syntactically enhanced and aspect-oriented multimodal aspect-level sentiment analysis model, the problem of multimodal data fusion was solved, enabling accurate analysis of different aspects of sentiment in user comments on online social platforms, and improving the accuracy and robustness of sentiment analysis.

CN122310083APending Publication Date: 2026-06-30HENAN UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-21
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Existing technologies struggle to effectively integrate multimodal data, especially in analyzing the sentiment tendencies of different aspects in user comments on online social platforms. Traditional methods are unable to accurately identify users' specific sentiment tendencies towards various aspects.

Method used

We construct a multimodal aspect-level sentiment analysis model based on syntactic enhancement and aspect-oriented approaches. Through feature extraction, syntactic enhancement, visual enhancement, multi-level cross-modal interaction, and auxiliary reconstruction modules, we achieve deep collaborative modeling and sentiment classification of multimodal data.

Benefits of technology

It significantly improves the accuracy and robustness of the model in aspect-level multimodal sentiment analysis, and can accurately identify the sentiment polarity of different aspects in scenarios with inconsistent text and images and visual noise.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122310083A_ABST
    Figure CN122310083A_ABST
Patent Text Reader

Abstract

This application relates to a multimodal aspect-level sentiment analysis method based on syntactic enhancement and aspect-oriented approaches. The method includes acquiring a multimodal dataset and preprocessing it to obtain a preprocessed dataset; constructing an aspect-level multimodal sentiment analysis model based on syntactic enhancement and aspect-oriented networks, and training it on the preprocessed multimodal dataset to obtain a trained aspect-level multimodal sentiment analysis model; acquiring the multimodal data to be analyzed and inputting it into the trained aspect-level multimodal sentiment analysis model; and obtaining the sentiment analysis results after processing by the model. This application, through the collaborative work of multiple modules such as syntactic enhancement and visual enhancement, can accurately identify the sentiment polarity of different aspects and maintains strong analytical performance even in scenarios with inconsistencies between text and images and visual noise.
Need to check novelty before this filing date? Find Prior Art