.

ISSN 2063-5346
For urgent queries please contact : +918130348310

PROCESS OF TEXT SENTIMENT ANALYSIS USING DEEP LEARNING FOR DOUBLE NEGATION ON PRODUCT REVIEWS

Main Article Content

Ramandeep Sandhu , Abhishek Yadav , Kartikey Sahu , Nitin Jaiswal , Piyush Pushkar , Mohammad Faiz
» doi: 10.31838/ecb/2023.12.s1-B.203

Abstract

In the world of E-commerce and customer interest in giving their sentiments online is increasing at fast pace. Sentiments can be of three forms such as positive, negative or even sometimes sentiment in neutral form. On the other side, text-opinions which are given by customer using their own words choice is really a great challenge for reputation of a product company. In this paper, the proposed approach is working on analysis of double negation which is actually residing in negative comments but the real meaning of such comments is towards high rank of product as well as for company which is dealing with that product. A novel approach using a Long Short-Term Memory (LSTM) neural network for end-to-end sentiment analysis along with the inclusion of negation identification is implemented in this study. Our approach uses the word embedding to transform the text data into a numerical format, and introduces a Hyperparameter tuning, performed using a grid search approach to find the optimal combination of hyperparameters for the EmbedLSTM model. The approach uses a dataset obtained from Kaggle with sentiments having negation at double level. The performance of the model is evaluated using various metrics such as accuracy, precision, recall, F1 score, and specificity. Our methodology provides a rigorous approach to developing a sentiment analysis model that can be applied to a variety of datasets.

Article Details