A Pipeline Approach to Research Paper Recommendation System Integrating TF-IDF and KNN

The literature review is a critical analysis of existing research papers, systematically summarizing and evaluating relevant studies, books, and sources to offer a comprehensive overview and identify gaps in knowledge. It guides researchers in formulating questions and contributes to the existing body of knowledge. However, researchers often struggle to locate suitable articles, especially those with limited search expertise. Traditional recommendation methods sometimes don't work well in the ever-changing world of academic papers. It's tough to keep up with the newest research and find high quality papers that are less cited. Recognizing these limitations, the proposed research paper recommendation system stands out as an innovative solution to help researchers overcome difficulties. At its core, the system employs a sophisticated pipelined hybrid approach, combining content-based strategies with collaborative filtering techniques. This integration aims to overcome the shortcomings of conventional methods by providing a more robust recommendation mechanism. Content-based methods include TF-IDF with cosine similarity. Collaborative filtering includes Knn, seamlessly blending content-based insights into collaborative filtering results. The system's dataset is sourced from ArXiv, which has provided open access to scholarly articles for almost 30 years. Spanning various fields, this extensive repository, available through Kaggle, encompasses 1.7 million articles with features like titles, authors, categories, abstracts, and full text PDFs. It offers significant depth for efficient data extraction. By blending the strengths of arXiv dataset, content-based and collaborative filtering, the system achieves a synergy that not only captures the richness of individual methods but also harnesses their combined power to deliver recommendations that are both timely and comprehensive.