<![CDATA[This study classifies Korean dramas based on popularity and ratings using K-Means Clustering within the CRISP-DM framework. The dataset from Kaggle includes title, release year, rating, vote count, duration, and genre. The Elbow Method determined that 2 clusters were optimal, with a Silhouette Score of 0.35, indicating a fairly good grouping. Recommended dramas have high ratings and vote counts, showing strong popularity, while less recommended dramas have lower ratings and fewer votes, indicating limited appeal. This model can enhance recommendation system accuracy, assist viewers in content selection, and help streaming platforms understand user trends and marketing strategies. Future improvements may involve alternative clustering methods (DBSCAN, Hierarchical Clustering) and additional features like actors, directors, and release year to refine accuracy.]]>
Copyrights © 2025
