Smart content aggregator: Leveraging NLP and machine learning for real-time personalized information

This study introduces a Smart Content Aggregator that uses NLP and machine learning to deliver real-time, personalized information by collecting and summarizing content from sources like news websites, blogs, and social media. The system applies multiple NLP-based feature extraction methods TF-IDF, Word2Vec with cosine similarity, and Sentence-BERT to filter and rank content based on user-defined preferences. Comparative analysis shows that Sentence-BERT achieves the highest relevance accuracy (94.3%) and semantic match quality, though at the cost of higher response time 5.6s. Meanwhile, TF-IDF provides faster results 2.1s with moderate accuracy 83.4%. Overall, the proposed NLP-ML Enhanced Model improves over traditional baselines with a 23.5% gain in F1-score, 16.1% higher precision, and 15.2% better recall, making it effective for reducing noise and delivering relevant content to users. These improvements highlight the system's effectiveness in reducing redundant information and delivering relevant content. The aggregator can be applied in areas like education, journalism, research, and personalized news delivery, where timely and accurate information is critical.