Feature Selection based Sentiment analysis using Combination of Nature Inspired Algorithms

Social media now-a-days in the best source of collecting information about any product, movie and many more. The persons using the social media prefer to share their reviews or comments in different platforms like X, Facebook, online shopping sites, which helps others to get information about the product. Even company people, who developed the product also helpful by these reviews to know the shortcomings of their product and able to prepare a new product after eliminating the issues. These reviews can be collected from the sources and can be processed to generate information. These reviews need to go through the different process until the final conclusion is obtained. When the dataset containing the reviews are large in size, processing time is increased and thus, feature selection technique is preferred, which will select a representation of the features. Thus, in this work, a combination of two nature inspired algorithms are considered i.e., Cat Swarm Optimization (CSO) and Artificial Neural Network (ANN). The CWO acts as the feature selection algorithm and the selected features are given input to ANN for classification. Finally, the result is calculated with the help of confusion matrix and other performance evaluation parameters associated with it.