Twitter Data Sentiment Analysis Model: GuianSpin-Convolutional Network

Abstract

Millions of people use microblogging to share their thoughts on a variety of subjects and to get feedback on any product, service, or problem. Sentiment analysis is the process of recognizing and categorizing the sentiments conveyed in the source text. Twitter is a well-known microblogging platform where monitoring and summarizing various thinking processes can yield valuable insights into various points of view. Thus, a major difficulty in the current research is to analyses opinions and categories them based on their polarity (positive or negative). In order to improve the effectiveness of sentiment evaluation in addition to other textual analysis, the GuianSpin-Convolutional Network is used in this study to predict the emotions from the tweets. This network combines GuianSpin optimisation with CNN.The TF-IDF technique, which improves sentiment analysis, is used to extract contextual and semantic information from the Twitter data. To get beyond the drawbacks and difficulties associated with feature extraction and data classification when utilising the GuainSpin Convolutional Network, extracted features are chosen and then applied to GuianSpin optimisation, which has the capacity to randomly create the solution and sequentially optimise. By altering the network parameters according to the fitness estimation, the GuianSpin enhances CNN training by making it easier to extract features from text data without changing the structure of the model. It improves the CNN's rate of convergence, optimises sentiment analysis solutions, and adapts dynamically to the context and degree of difficulty.The model achieves  high accuracy , sensitivity  and  specificity of even while considering the TP utilizing the Twitter sentiment database.