Sentiment analysis (SA) is a necessary duty for various areas where this
method is very important to know public opinion of users about brands,
products, events, etc. In this way, customer satisfaction can be
increased for companies by making the necessary changes on the product.
In this study, we aim to implement a SA on the comments made by users on
online sales sites. We propose and test four machine learning (ML)
algorithms and a deep learning (DL) model. Feature selection methods
enable algorithms to capture existing patterns more easily and reduce
running times by finding features that contribute highly to
classification within the dataset and reducing the search space.
Therefore, we apply the binary version of Sailfish Optimization (SOA),
also called Binary Sailfish Optimizer (BSO), as a feature selector to a
textual dataset and run for SA for the first time. In order to evaluate
its performance, we make a comparison with two optimization algorithms,
named Harmony Search (HS) and Bat Algorithm (BA). The results show that
the BSO is ahead of the HS and BA algorithms with an F-Score of 0.91,
using almost half of the available features.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.