Sign in Subscribe
Machine learning

Heuristic Multi-Scale Feature Fusion with Attention-Based CNN Achieves Higher Accuracy in Sentiment Analysis

Strategic Insights

— 1 min read

A new study on neural computation has made a breakthrough in sentiment analysis using a proposed framework that combines deep learning and heuristic methods. The research, conducted by the Department of Computer Science and Engineering at RMK College of Engineering and Technology in Thiruvallur, India, aimed to overcome the limitations of traditional sentiment analysis methods by leveraging the strengths of attention-based Convolutional Neural Networks (CNNs) and multi-scale feature fusion.

Key Takeaways:

  • The proposed framework uses an attention deep learning model with an improved heuristic approach to extract features from input text data, achieving higher accuracy compared to traditional methods.
  • The input text data is gathered from public resources and pre-processed to prevent unrelated text data, followed by feature extraction using Bidirectional Encoder Representations from Transformers (BERT), Transformers, and word2vector.
  • The resultant features are fused and subjected to a Multiscale Feature Fusion-based Adaptive and Attention-based Convolution Neural Network (MFF-AACNet), where the sentiment is analyzed using an improved Fitness Opposition of Rat Swarm Optimizer (FORSO) for parameter tuning.
  • The proposed framework demonstrates a significant improvement in sentiment analysis accuracy, with the ability to navigate through datasets with varying scales and complexities.

Statistics:

  • The research concludes that the proposed framework achieves higher accuracy compared to traditional methods, indicating a significant improvement in sentiment analysis.
  • The authors report that their model achieves a performance of 92% accuracy in sentiment analysis, outperforming traditional methods.
  • The study highlights the potential of the proposed framework in real-world applications, including customer service and social media analysis, where accurate sentiment analysis is crucial.

Sources:

  • Heuristic multi-scale feature fusion with attention-based CNN for sentiment analysis. Network-computation In Neural Systems, 2025:1-41.
  • Department of Computer Science and Engineering Reports Findings in Neural Computation (Heuristic multi-scale feature fusion with attention-based CNN for sentiment analysis). Robotics & Machine Learning. May 19, 2025; p 123.

Read more