Intellectual Property India Publishes Patent Application for Artificial Intelligence-Based Framework For Customer Query Analysis Using Domain-Specific Embedded Clustering And Ensembled Classification

MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641086373 A) filed by Dr. C. Kiruthiga; and Dr. K. Dharmarajan on July 15, 2026, for Artificial Intelligence-Based Framework For Customer Query Analysis Using Domain-Specific Embedded Clustering And Ensembled Classification.

Inventors include Dr. C. Kiruthiga; and Dr. K. Dharmarajan.

The application for the patent was published on August 28, 2026, under issue no. 35/2026.

Abstract: The present invention relates to an Artificial Intelligence-based framework for customer query analysis using domain-specific embedded clustering and ensemble classification techniques. The framework is designed to process large volumes of unstructured customer queries obtained from sources such as call centers and digital platforms. It incorporates a multi-phase architecture consisting of preprocessing, semantic embedding, clustering, classification, and validation modules. The preprocessing phase includes exploratory data analysis, data and feature refinement, feature selection, and domain-specific auto spell correction to enhance data quality. The system utilizes advanced embedding techniques such as Word2Vec and Sentence-BERT to capture contextual and semantic relationships within queries. A hybrid clustering approach, including HDBSCAN and Deep Embedded Clustering (DEC), is employed to group semantically similar queries effectively. The classification phase integrates machine learning and deep learning models, including Naive Bayes and Long Short-Term Memory (LSTM), along with an ensemble classification mechanism to improve prediction accuracy and robustness. The framework also includes a performance evaluation module that measures accuracy, precision, recall, F1-score, and query processing time to ensure efficiency and reliability. The invention supports real-time query processing and scalable deployment across multiple domains. Overall, the proposed system enhances customer interaction by providing accurate, efficient, and context-aware query analysis and response generation. FIG.1.

Disclaimer: Curated by HT Syndication.