Indisent: Cross-lingual sentiment analysis

We propose Indisent, an api service that provides sentiment analysis for 11 indic languages and is also capable of translating text from indic languages to english and vice versa. Indisent performs translation by making use of neural machine translation based on a transformer model trained on the sa...

Full description

Saved in:
Bibliographic Details
Main Authors: Ansari, Nazneen, lopes, Monalisa, Shaikh, Hussain, D’Mello, Blaise, Shah, Dharmit, Rodrigues, Linson
Format: Conference Proceeding
Language:English
Subjects:
Languages
Machine translation
Sentiment analysis
Translating
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose Indisent, an api service that provides sentiment analysis for 11 indic languages and is also capable of translating text from indic languages to english and vice versa. Indisent performs translation by making use of neural machine translation based on a transformer model trained on the samanantar dataset. It performs sentiment analysis by translating text into english and then uses a pre-trained sentiment analysis model SiEBERT. Using this approach we are able to achieve cross-lingual sentiment analysis with increased accuracy for low resource indic languages. Using state of the art transformer models the need of training multiple models for each language is also eliminated.
ISSN:0094-243X
1551-7616
DOI:10.1063/5.0240373