<![CDATA[Examining consumer evaluations of food on social media provides relevant in- formation for anyone searching, especially immigrants and tourists. This infor- mation is also highly valuable for food stall owners and restaurant managers be- cause it helps them improve the quality of the food they serve based on customer feedback. However, sentiment analysis of food reviews often faces challenges due to inadequate data preprocessing, which leads to low classification accuracy. This study aims to improve sentiment recognition accuracy in food reviews by optimizing the feature attribute selection process in the classification model. The classification model employed in this research is Naive Bayes (NB), enhanced through a hybrid feature selection approach that combines the information gain (IG) algorithm and the genetic algorithm (GA). This combination is designed to maximize the selection of the most relevant feature attributes, thereby improving the model’s ability to identify positive, negative, and neutral sentiments in con- sumer food reviews on social media. The experimental results show that the hybrid IG-GA model achieved the highest accuracy rate of 93%, outperform- ing models that use individual algorithms. These findings demonstrate that the hybrid feature selection method effectively enhances the sentiment analysis performance of the Naive Bayes model. This study contributes to the develop- ment of food recommendation systems, the improvement of service strategies for culinary businesses, and supports the achievement of SDG 8 (Decent Work and Economic Growth) and SDG 9 (Industry, Innovation, and Infrastructure).]]>
Copyrights © 2025
