Topic Modeling for Research Paper Categorization

Abstract

"The evolving of academic research domain is raising rapidly, with a spreading increase in the volume and diversity of scholarly publications across vast various of domains. Due to this huge increment of academic research domain, the stake holders such as researchers, educators and knowledge seekers have faced a difficulty to navigate their research in the correct path in this vast ocean of information to find the relevant research papers effectively. The problem lies in the need for an innovative system that can streamline the organization and accessibility of academic literature. Traditionally this process has been time consuming and ponderous. In this project it will aim to provide a solution to this challenge, using advanced topic modeling and text classification techniques. This will ease the research papers search process and will do it effectively. Building upon the Natural Language Processing (NLP) and deep learning, this system seeks to revolutionize the way researchers’ access and categorize research papers. Through this project the main purpose is to create a robust system that automatically categorizes research papers into relevant topics and themes. This categorization will help users to navigate through a collection of academic literature effortlessly. to overcome this challenge, it is proposing a Latent Dirichlet Allocation (LDA) model to categorize research papers. Then when a researcher search for a research paper through this system it will show only the relevant research papers to the searched domain. In the present research papers search systems using the key word search for the search process. As per the google the google scholar (Google Scholar, 2019) uses the key word search to produce results. But through this system it will not search for only key words, it’ll show results based on the domain also. "