<![CDATA[This study presents a comprehensive analysis of sentiment classification algorithms applied to content from the entertainment industry, specifically focusing on hip-hop music videos. Following the CRISP-DM (Cross-Industry Standard Process for Data Mining) methodology, the research evaluates the performance of three prominent algorithms: k-nearest Neighbors (k-NN), Decision Tree (DT), and Support Vector Machine (SVM) with Synthetic Minority Over-sampling Technique (SMOTE). The analysis incorporates performance metrics, including accuracy, precision, recall, f-measure, and the area under the curve (AUC) values. The dataset comprises user-generated comments and feedback from two distinct hip-hop music videos. Results indicate that all three algorithms exhibit notable accuracy in classifying sentiments, with SVM with SMOTE achieving the highest accuracy of 83.68%. DT demonstrates balanced performance metrics, particularly in precision and recall, with an accuracy of 79.12%. Meanwhile, k-NN exhibits a lower accuracy of 64.71% but showcases balanced precision and recall rates. These findings suggest the suitability of SVM with SMOTE for sentiment classification tasks in the entertainment industry, offering valuable insights for content creators, marketers, and platform administrators to enhance audience engagement and user experience. Additionally, the study underscores the importance of algorithmic evaluation and selection in content analysis, providing guidance for future research and practical applications in the entertainment domain within the framework of CRISP-DM.]]>
Copyrights © 2024
