HGATT_LR: transforming review text classification with hypergraphs attention layer and logistic regression.

Area of Science:

• Natural Language Processing
• Machine Learning
• Data Mining

Background:

• Text classification is crucial for sentiment analysis, opinion mining, and customer feedback.
• Hypergraph algorithms effectively capture complex relationships within text data.
• Existing methods require improvement for intricate real-world datasets.

Purpose of the Study:

• To propose a novel Hypergraph Attention Layer with Logistic Regression (HGATT_LR) model for enhanced text classification.
• To evaluate the performance of HGATT_LR on the Amazon review dataset.
• To demonstrate the superiority of hypergraph approaches in handling complex text interactions.

Main Methods:

• Text preprocessing, keyword extraction using Latent Dirichlet Allocation (LDA), and feature selection using node-level and edge-level attention.
• Development of a Hypergraph Attention Layer integrated with Logistic Regression (HGATT_LR).
• Comparative analysis against state-of-the-art text classifiers on the Amazon review dataset.

Main Results:

• The proposed HGATT_LR model achieved 88% accuracy in text classification.
• HGATT_LR demonstrated superior performance compared to other evaluated text classification algorithms.
• The model proved scalable and adaptable to larger datasets.

Conclusions:

• Hypergraph Attention Networks offer a powerful approach for text classification, particularly for datasets with complex interdependencies.
• The HGATT_LR model provides a scalable and effective solution for real-world text analysis.
• This research empowers businesses to improve product quality through advanced text analysis.