Intellectual Property India Publishes Patent Application for 'Deep Learning Approaches Cnn And Rnns To Sentiment Analysis In Social Media With Word Embeddings' Filed by Mr. Nagaraju Pamarthi; Mr. Siva Yenikepalli; Mr. Somanagouda Patil; Rupangudi Saila Banu; Dr. Deekonda Anusha; Esai Yazhini P; Mr. Veer Sudheer Goud; and Prof. Neha Singh

MUMBAI, India, Feb. 13 -- Intellectual Property India has published a patent application (202641008747 A) filed by Mr. Nagaraju Pamarthi; Mr. Siva Yenikepalli; Mr. Somanagouda Patil; Rupangudi Saila Banu; Dr. Deekonda Anusha; Esai Yazhini P; Mr. Veer Sudheer Goud; and Prof. Neha Singh, Kovvada, Andhra Pradesh, on Jan. 28, for 'deep learning approaches cnn and rnns to sentiment analysis in social media with word embeddings.'

Inventor(s) include Mr. Nagaraju Pamarthi; Mr. Siva Yenikepalli; Mr. Somanagouda Patil; Rupangudi Saila Banu; Dr. Deekonda Anusha; Esai Yazhini P; Mr. Veer Sudheer Goud; and Prof. Neha Singh.

The application for the patent was published on Feb. 13, under issue no. 07/2026.

According to the abstract released by the Intellectual Property India: "The invention presents an advanced system and method for sentiment analysis of social media content using deep learning techniques integrated with semantic word embedding representations. With the rapid growth of social media platforms, vast amounts of user-generated textual data are continuously produced, reflecting public opinions, emotions, and attitudes across diverse topics. Traditional sentiment analysis approaches, which rely on rule-based methods or conventional machine learning algorithms with handcrafted features, often fail to capture contextual meaning, semantic relationships, and linguistic variability inherent in social media text. The proposed invention overcomes these limitations by employing a deep learning framework that combines convolutional neural networks and recurrent neural networks for automated and accurate sentiment classification. Convolutional neural networks are utilized to identify local sentiment-bearing patterns and key phrases within text, while recurrent neural networks, including long short-term memory and gated recurrent unit architectures, model sequential dependencies and contextual relationships across entire text sequences. This combined approach enables the system to capture both local and global features essential for precise sentiment interpretation. A core component of the invention is the use of word embeddings to represent textual data as dense, low-dimensional vectors that encode semantic and syntactic information. By leveraging embedding techniques such as Word2Vec, GloVe, or fastText, the system can generalize across different expressions of sentiment, including informal language, slang, abbreviations, and evolving vocabulary commonly found in social media. This enhances robustness and adaptability compared to traditional sparse text representations. The invention is designed to support large-scale and real-time sentiment analysis applications, providing efficient processing of high-volume social media data with high classification accuracy. The system is flexible and can be adapted for binary or multi-class sentiment classification across various domains and platforms. Overall, the invention delivers a scalable, accurate, and intelligent sentiment analysis solution, offering significant improvements over existing techniques and enabling effective opinion mining, trend analysis, and decision support in social media environments."

Disclaimer: Curated by HT Syndication.