BERTMelon

Sentiment analysis using BERT on Hindi-English code-mixed data.

Introduction

According to Wikipedia -

Sentiment analysis (also known as opinion mining or emotion AI) refers to the use of natural language processing, text analysis, computational linguistics, and biometrics to systematically identify, extract, quantify, and study affective states and subjective information.

Task

To predict the sentiment of code-mixed Hindi-English data with high accuracy.

TODO

1. Use talos to get better hyperparameter values.
2. Try the model on a different code-mixed dataset.