Abstract: TheInternet is widely used for the propagation of propaganda as well as thesharing of thoughts and points of view. This research proposes the use ofsentiment analysis techniques to decode web forum posts in a variety oflanguages. The article examines the usefulness of stylistic and syntacticfeatures in classifying emotions in English and Arabic language. To account forArabic linguistic features, unique feature extraction elements have been added.Additionally, the Entropy Weighted Genetic Algorithm (EWGA) is developed, whichis a hybridized genetic algorithm that uses the information gain heuristic toselect features. EWGA's objective is to improve continuity and to provide moreaccurate evaluation of critical functions. The suggested features and methodsare tested using baseline movie review data from the United States and theMiddle East, as well as web site postings from other countries. Theexperimental results obtained by integrating EWGA and SVM show that the deviceworks well, with an accuracy of over 95% on the benchmark data collection andover 93% on both the US and Middle Eastern forums. Stylistic featuressignificantly increased results in all test beds, and EWGA outperformedalternative feature selection approaches, demonstrating the efficacy of thesefeatures and techniques for document level sentiment classification.
.Keywords-Competitorsfeatures, Hotel, E-Commerce, Seller Competitors