MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641076885 A) filed by Sr University on June 22, 2026, for A Hybrid System For Transparent And Accurate Sentiment Prediction On Short-Form Social Media Streams And Method Thereof.
Inventors include Nikita Gaur; and Dr. Sridhar Chintala.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: A HYBRID SYSTEM FOR TRANSPARENT AND ACCURATE SENTIMENT PREDICTION ON SHORT-FORM SOCIAL MEDIA STREAMS AND METHOD THEREOF A hybrid approach to sentiment analysis on Twitter provides enhanced classification accuracy and model interpretability by integrating Natural Language Processing (NLP) with an Explainable Artificial Intelligence (XAI) verification tier. The system features a Multi-Modal Feature Extraction Pipeline that processes raw Twitter data to extract linguistic structures including bag-of-words, TF-IDF, sentiment-based elements, and punctuation patterns. A Feature Selection Engine executes an integrated metaheuristic routine combining the Golf Optimization Algorithm (GOA) and the Serval Optimization Algorithm (SOA) to discard redundant dimensions. The optimization results are processed by a core HAConv-Net Classifier, which merges a Hierarchical Attention Network (HAN) with a Siamese Convolutional Neural Network (SCNN) to categorize sentiments into positive, negative, or neutral expressions. Post-classification, an XAI module applies game-theoretic SHAP or LIME frameworks to render interactive localized feature attribution summaries, enabling transparent, legally justifiable verification of textual sentiment assessments suitable for real-time brand management and public health monitoring.
Disclaimer: Curated by HT Syndication.
