Aspect Category Detection Using Bi-LSTM-Based Deep Learning for Sentiment Analysis for Hindi

Abstract

In the field of sentiment analysis, the notion of aspect category identification lays emphasis on identifying the aspect categories in a specific review phrase. The purpose of this research is to propose a novel approach to the identification of aspect categories in reviews that are published in Hindi. The training and evaluation of a supervised model that is based on deep learning enable the extraction of aspect categories. Every experiment is carried out with a well-accepted Hindi dataset. One of the challenges that are involved in aspect category recognition is the classification of text that has several labels. The utilization of a deep neural network that is founded on BiLSTM leads to an enhancement of the category detection outcomes. The results came out with an F-score of 0.8345 and an accuracy of 93.91% when applied to the well-known Hindi dataset. The offered architecture, in conjunction with the results that were achieved, gives a great deal of significance because it serves as a fundamental resource for future research and activities related to the issue. In this study, a deep-learning architecture is proposed for the aim of detecting aspect categories in Hindi. The outcomes of this architecture are both new and state-of-the-art in their respective fields.