Paper Key : IRJ************140
Author: Shubham Vaibhav Kanse,Onkar Anandrao Desai,Dr. Urmila R Pol
Date Published: 01 Apr 2025
Abstract
In this paper, we used deep learning techniques to investigate the process of emotion analysis of IMDb movie reviews. Emotional analysis, which ascertains the text input's emotional tone, is a crucial part of natural language processing (NLP). The study's objective is to categorize IMDB movie reviews as either positive or negative using deep learning models like long short-term memory (LSTM) and recurrent nervous network (RNN). The dataset is pre-developed utilizing techniques including lemmatization, steaming, tokenization, and stop-ward removal in order to improve the model's performance. Text data is converted into numerical format using feature extraction techniques such as Word embedding and Term Frequency Document Frequency (TF -DF). Major performance indicators such as accuracy, accuracy, recall, and F1 scores are used to train and assess the model. Conclusions suggest that in terms of overall performance and future accuracy, especially deep learning models, LSTM performs better than traditional machine learning models. This study further enhances automatic emotion analysis by enhancing the movie recommendation system and providing practical information to understand the viewer's input
DOI Requested