Content Recommendation Based on Topic Modeling

Abstract

With the proliferation in Internet usage and communicating devices, plenty amount of information is available at user disposal but on other side, it leads to a challenge to provide the fruitful information to end users. To overcome this problem, recommendation system plays a decisive role in providing pragmatic information to end users at appropriate time. This paper proposes a topic modeling based recommendation system to provide contents related to end users interest. Recommendation systems are based on different filtering mechanisms which are classified as content based, collaborative based, knowledge based, utility based and hybrid filtering, etc. The objective of this research is thus to proffer a recommendation system based on topic modeling. Benefit of latent Dirichlet allocation (LDA) is to uncover latent semantic structure from the text documents. By analyzing the contents using topic modeling, system can recommend the right articles to end users based on user interest.

Apples to Apples: A Systematic Evaluation of Topic Models

“…It has the advantage of not relying on any labeled data to achieve good results, as the training of topic models is done in an unsupervised matter. Moreover, the resulting topics and representations can then be used to perform other NLP tasks such as trend prediction (Lau et al, 2012), text summarization (Lin and Hovy, 2000), improving named entity recognition (Newman et al, 2006), and content recommendation (Papneja et al, 2021).…”