Multi-Label Research Paper Classification and Recommendationsystem

In the recent years, due to the increasing number of research papers it has become difficult for researchers to keep up with the development in their fields. However, the current solutions for this problem are limited as they either support single label classification and low efficiency in multi-label classification. To overcome this limitation, this paper proposes a multi-label research paper classification and recommendation system. It addresses problems of preprocessing, handling data imbalance issues, and employs the LSTM for the multi-label classification task. The experimentation conducted on “Arxiv paper abstract dataset” and has obtained effective accuracy of 98.35%. The model proposed performs better than the current models in the area.