Analysis of Customer Sentiment towards Wardah Beauty Products on Instagram Using K-Nearest Neighbors (KNN) and Naïve Bayes Algorithms

The beauty industry in Indonesia continues to grow rapidly, with Wardah emerging as a leading local brand that garners numerous customer reviews on Instagram. Understanding these customer opinions is crucial, as they significantly influence brand perception and purchasing decisions. This study aims to compare the performance of K-Nearest Neighbors (KNN) and Naïve Bayes algorithms in classifying customer sentiment toward Wardah’s products on Instagram. The research follows the CRISP-DM methodology, involving data collection through web scraping, data preprocessing, and classification using both algorithms. The results show that KNN achieves higher accuracy (78.23%) than Naïve Bayes (74.56%), making it more effective in sentiment classification. KNN is advantageous in identifying distance-based patterns, while Naïve Bayes offers faster processing and easier implementation. These findings suggest that KNN is more suitable for accurate sentiment prediction. Wardah can implement a KNN-based sentiment monitoring system to track public feedback on product launches and campaigns, thereby enhancing marketing strategies and customer engagement