Spider Monkey Optimization with Deep Learning-based Hindi Short Text Sentiment Analysis

 

Praloy Biswas1,*, A. Daniel2, Subhrendu Guha Neogi3

 

1 Department of Computer Science and Engineering, Amity University Madhya Pradesh, Gwalior, India

2 Department of Computer Science and Engineering, Amity University Madhya Pradesh, Gwalior, India

3 Department of Computer Science and Engineering, The Neotia University, West Bengal, Kolkata, India 

Emails: praloy.biswas.it@gmail.com;  adaniel@gwa.amity.edu;  sguhaneogi@gmail.com

*Corresponding Author: praloy.biswas.it@gmail.com

 

Text Box: Abstract Sentiment analysis (SA) intends to categorize a text respective to sentimental polarity of individual opinions, like neutral, positive, or negative. The study of Hindi is limited because of the grammatical and morphological complexities of the Hindi language while many research work concentrates on drawing features from English text. The hindi languages make the sentiment classification procedure for Hindi short text a tedious process. The Hindi language has complicated morphology and variation based on phonetics, spelling, and vocabulary; the common usage of numerous dialects between Hindi in India produces a massive volume of glossaries. In this study, we introduce a Spider Monkey Optimization with stacked recurrent neural network (SMO-SRNN) for short text SA on Hindi Corpus. The proposed SMO-SRNN technique mainly aims to identify and categorize the Hindi short text into three distinct classes, namely negative, positive, and neutral. In the presented SMO-SRNN method, the SRNN approach is exploited for the investigation and classification of sentiment. Moreover, the SMO model is employed to finetune the hyperparameter related to the SRNN model. A detailed set of experiments is applied to ensure the high efficiency of the SMO-SRNN algorithm. The comparative outcome highlighted the enhancement of the SMO-SRNN technique over other methods.  Received: August 16, 2023  Revised: November 09, 2023  Accepted: February 27, 2024

 

Keywords: Short text classification; sentiment analysis; deep learning; spider monkey optimization; recurrent neural network