Hybrid BPSO-IGA Approach for Semantic Summarization of Hindi Documents

The rapid increase in digital Hindi news and online textual content has created a strong need for intelligent systems that can automatically generate meaningful summaries. This project, titled Hybrid BPSO IGA Based Semantic Summarization System for Hindi Documents, presents an efficient approach to reduce large Hindi texts into concise and informative summaries. The proposed system uses a hybrid optimization technique combining Binary Particle Swarm Optimization and an Improved Genetic Algorithm to select the most relevant and non repetitive sentences. Linguistic and statistical features such as TF IDF score, sentence length, lexical diversity, and keyword relevance are enhanced with multilingual semantic sentence embeddings to better understand contextual meaning. Maximal Marginal Relevance is applied to maintain diversity in the selected content. The extracted summary is further refined using the multilingual mT5 transformer model to improve fluency and readability. The results demonstrate that the system produces accurate and coherent summaries with effective compression, making it suitable for real world applications involving large scale Hindi text data.