Aspect-Based Sentiment Analysis in Hindi Language by Ensembling Pre-Trained mBERT Models

Sentiment Analysis is becoming an essential task for academics, as well as for commercial companies. However, most current approaches only identify the overall polarity of a sentence, instead of the polarity of each aspect mentioned in the sentence. Aspect-Based Sentiment Analysis (ABSA) identifies...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Abhilash Pathak, Sudhanshu Kumar, Partha Pratim Roy, Byung-Gyu Kim
Formato: article
Lenguaje:EN
Publicado: MDPI AG 2021
Materias:
Acceso en línea:https://doaj.org/article/f31531bbf9094a2ab04c2351944bae2e
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:f31531bbf9094a2ab04c2351944bae2e
record_format dspace
spelling oai:doaj.org-article:f31531bbf9094a2ab04c2351944bae2e2021-11-11T15:39:03ZAspect-Based Sentiment Analysis in Hindi Language by Ensembling Pre-Trained mBERT Models10.3390/electronics102126412079-9292https://doaj.org/article/f31531bbf9094a2ab04c2351944bae2e2021-10-01T00:00:00Zhttps://www.mdpi.com/2079-9292/10/21/2641https://doaj.org/toc/2079-9292Sentiment Analysis is becoming an essential task for academics, as well as for commercial companies. However, most current approaches only identify the overall polarity of a sentence, instead of the polarity of each aspect mentioned in the sentence. Aspect-Based Sentiment Analysis (ABSA) identifies the aspects within the given sentence, and the sentiment that was expressed for each aspect. Recently, the use of pre-trained models such as BERT has achieved state-of-the-art results in the field of natural language processing. In this paper, we propose two ensemble models based on multilingual-BERT, namely, <i>mBERT-E-MV</i> and <i>mBERT-E-AS</i>. Using different methods, we construct an auxiliary sentence from this aspect and convert the ABSA problem to a sentence-pair classification task. We then fine-tune different pre-trained BERT models and ensemble them for a final prediction based on the proposed model; we achieve new, state-of-the-art results for datasets belonging to different domains in the Hindi language.Abhilash PathakSudhanshu KumarPartha Pratim RoyByung-Gyu KimMDPI AGarticleaspect-based sentiment analysisBERTclassificationensembleHindiElectronicsTK7800-8360ENElectronics, Vol 10, Iss 2641, p 2641 (2021)
institution DOAJ
collection DOAJ
language EN
topic aspect-based sentiment analysis
BERT
classification
ensemble
Hindi
Electronics
TK7800-8360
spellingShingle aspect-based sentiment analysis
BERT
classification
ensemble
Hindi
Electronics
TK7800-8360
Abhilash Pathak
Sudhanshu Kumar
Partha Pratim Roy
Byung-Gyu Kim
Aspect-Based Sentiment Analysis in Hindi Language by Ensembling Pre-Trained mBERT Models
description Sentiment Analysis is becoming an essential task for academics, as well as for commercial companies. However, most current approaches only identify the overall polarity of a sentence, instead of the polarity of each aspect mentioned in the sentence. Aspect-Based Sentiment Analysis (ABSA) identifies the aspects within the given sentence, and the sentiment that was expressed for each aspect. Recently, the use of pre-trained models such as BERT has achieved state-of-the-art results in the field of natural language processing. In this paper, we propose two ensemble models based on multilingual-BERT, namely, <i>mBERT-E-MV</i> and <i>mBERT-E-AS</i>. Using different methods, we construct an auxiliary sentence from this aspect and convert the ABSA problem to a sentence-pair classification task. We then fine-tune different pre-trained BERT models and ensemble them for a final prediction based on the proposed model; we achieve new, state-of-the-art results for datasets belonging to different domains in the Hindi language.
format article
author Abhilash Pathak
Sudhanshu Kumar
Partha Pratim Roy
Byung-Gyu Kim
author_facet Abhilash Pathak
Sudhanshu Kumar
Partha Pratim Roy
Byung-Gyu Kim
author_sort Abhilash Pathak
title Aspect-Based Sentiment Analysis in Hindi Language by Ensembling Pre-Trained mBERT Models
title_short Aspect-Based Sentiment Analysis in Hindi Language by Ensembling Pre-Trained mBERT Models
title_full Aspect-Based Sentiment Analysis in Hindi Language by Ensembling Pre-Trained mBERT Models
title_fullStr Aspect-Based Sentiment Analysis in Hindi Language by Ensembling Pre-Trained mBERT Models
title_full_unstemmed Aspect-Based Sentiment Analysis in Hindi Language by Ensembling Pre-Trained mBERT Models
title_sort aspect-based sentiment analysis in hindi language by ensembling pre-trained mbert models
publisher MDPI AG
publishDate 2021
url https://doaj.org/article/f31531bbf9094a2ab04c2351944bae2e
work_keys_str_mv AT abhilashpathak aspectbasedsentimentanalysisinhindilanguagebyensemblingpretrainedmbertmodels
AT sudhanshukumar aspectbasedsentimentanalysisinhindilanguagebyensemblingpretrainedmbertmodels
AT parthapratimroy aspectbasedsentimentanalysisinhindilanguagebyensemblingpretrainedmbertmodels
AT byunggyukim aspectbasedsentimentanalysisinhindilanguagebyensemblingpretrainedmbertmodels
_version_ 1718434707024117760