Topic Modeling on Podcast Short-Text Metadata

TL;DR

This paper explores the use of topic modeling on podcast metadata, especially short texts, by incorporating named entities to improve topic coherence, with experiments on multiple datasets demonstrating the effectiveness of the proposed approach.

Contribution

It introduces NEiCE, a novel document representation leveraging named entities within a Non-negative Matrix Factorization framework for better topic modeling of podcast metadata.

Findings

NEiCE improves topic coherence over baselines

Experiments on datasets from Spotify, iTunes, and Deezer validate the approach

Proposed method enhances organization and navigation of podcast collections

Abstract

Podcasts have emerged as a massively consumed online content, notably due to wider accessibility of production means and scaled distribution through large streaming platforms. Categorization systems and information access technologies typically use topics as the primary way to organize or navigate podcast collections. However, annotating podcasts with topics is still quite problematic because the assigned editorial genres are broad, heterogeneous or misleading, or because of data challenges (e.g. short metadata text, noisy transcripts). Here, we assess the feasibility to discover relevant topics from podcast metadata, titles and descriptions, using topic modeling techniques for short text. We also propose a new strategy to leverage named entities (NEs), often present in podcast metadata, in a Non-negative Matrix Factorization (NMF) topic modeling framework. Our experiments on two existing datasets from Spotify and iTunes and Deezer, a new dataset from an online service providing a catalog of podcasts, show that our proposed document representation, NEiCE, leads to improved topic coherence over the baselines. We release the code for experimental reproducibility of the results.