<![CDATA[This research addresses the "rating-content discrepancy" on the Google Play Store, where numerical star ratings often conflict with the actual sentiment of textual reviews. Utilizing the CRISP-DM framework, the study evaluates the effectiveness of machine learning in resolving these inconsistencies by classifying Instagram user reviews into positive and negative categories. Two primary algorithms were compared using a dataset of 600 reviews. The Support Vector Machine (SVM) model demonstrated high efficacy with an accuracy of 0.84. In contrast, the K-Nearest Neighbors (KNN) model performed poorly, achieving an accuracy of only 0.48. This significant performance gap highlights SVM's superior ability to handle high-dimensional text data through optimal hyperplane separation. The research further integrated the Streamlit library to create an interactive web interface for real-time sentiment prediction and result visualization. Ultimately, this study confirms that a structured CRISP-DM approach combined with SVM provides a robust solution for automated opinion mining, offering a reliable methodology for future data science applications in social media analysis]]>
Copyrights © 2026
