Topic modeling and Recommendation system using LDA And Cosine Similarity

Topic modeling and Recommendation system using LDA And Cosine Similarity

Nayana J
Department of Machine Learning

B.M.S College Of Engineering

Bengaluru,India

nayana.ai21@bmsce.ac.in

Priyanka S
Department of Machine Learning

B.M.S College Of Engineering

Bengaluru,India

priyanka.ai21@bmsce.ac.in

Sowmya Lakshmi BS
Assistant Professor

Department of Machine Learning

B.M.S College Of Engineering

Bengaluru,India

sowmyalakshmibs.mel@bmsce.ac.in

Abstract
In this study, we explore the application of Natural Language Processing (NLP) techniques to analyze and recommend articles from Medium. By utilizing a dataset of Medium articles, we employ topic modeling using Latent Dirichlet Allocation (LDA) and implement a recommendation system using Doc2Vec to suggest articles similar to a given piece. The methodology includes preprocessing text data, visualizing insights, building an LDA model to discover latent topics, and creating a content-based recommendation system. Our findings show the potential of these techniques in enhancing user engagement and content discovery on digital platforms.

Index Terms- Content Recommendation, Latent Dirichlet Allocation, Natural Language Processing, Topic Modeling, User Engagement