# It's Only Words And Words Are All I Have

## TL;DR

This paper demonstrates that distributed representations of lyrics can effectively predict music genre and popularity, surpassing traditional handcrafted features and weak indicators, thus enhancing music mining and NLP tasks.

## Contribution

It introduces a novel approach using distributed lyric representations for genre and popularity prediction, overcoming limitations of prior handcrafted feature methods.

## Key findings

- Achieved around 0.6 F1 score for genre classification
- Achieved around 0.6 F1 score for popularity prediction
- Showed lyrics improve distributed representation quality

## Abstract

The central idea of this paper is to demonstrate the strength of lyrics for music mining and natural language processing (NLP) tasks using the distributed representation paradigm. For music mining, we address two prediction tasks for songs: genre and popularity. Existing works for both these problems have two major bottlenecks. First, they represent lyrics using handcrafted features that require intricate knowledge of language and music. Second, they consider lyrics as a weak indicator of genre and popularity. We overcome both the bottlenecks by representing lyrics using distributed representation. In our work, genre identification is a multi-class classification task whereas popularity prediction is a binary classification task. We achieve an F1 score of around 0.6 for both the tasks using only lyrics. Distributed representation of words is now heavily used for various NLP algorithms. We show that lyrics can be used to improve the quality of this representation.