<![CDATA[This research aims to classify public sentiment regarding the content of "Coral Reef 101," published by National Geographic. The methodology employed is the Cross-Industry Standard Process for Data Mining (CRISP-DM), encompassing stages such as business understanding, data understanding, modeling, evaluation, and deployment. The Decision Tree algorithm is utilized in conjunction with the SMOTE operator. This comprehensive approach enables the systematic analysis of public sentiment towards coral reef content, facilitating a deeper understanding of public perception and attitudes. The results of this study indicate that the DT algorithm with SMOTE demonstrates an accuracy of 87.51% +/- 4.28% (micro average: 87.50%), a precision of 80.35% +/- 5.10% (micro average: 80.00%) (positive class: Positive), recall of 100.00% +/- 0.00% (micro average: 100.00%) (positive class: Positive), f-measure of 89.02% +/- 3.22% (micro average: 88.89%) (positive class: Positive), and an AUC of 0.875 +/- 0.044 (micro average: 0.875) (positive class: Positive). These metrics demonstrate the effectiveness of the DT algorithm with SMOTE in accurately classifying public sentiment towards coral reef-related content, particularly in correctly identifying positive sentiment instances. The high accuracy, precision, recall, f-measure, and AUC values underscore the robustness and reliability of the model in sentiment analysis tasks.]]>
Copyrights © 2023
