DETERMINING OPTIMAL CLUSTERS AND DIMENSIONALITY REDUCTION FOR THE RAISIN DATASET USING K-MEANS CLUSTERING

Purpose Clustering algorithms play a crucial role in data mining by revealing hidden patterns in large datasets. This study focuses on identifying the optimal number of clusters and applying dimensionality reduction techniques to the raisin dataset using K-Means clustering. Design/Methodology/Approach The Elbow method is employed to determine the best value of k, enhancing K-Means performance. Dimensionality reduction techniques are applied to improve clustering interpretability and efficiency. K-Means is compared with DBSCAN and Hierarchical Clustering to evaluate clustering quality and computational efficiency. Findings Results indicate that K-Means, when combined with dimensionality reduction, outperforms DBSCAN and Hierarchical Clustering in terms of accuracy and computational time. Dimensionality reduction also enhances result visualization and interpretability. Originality/Value This study provides practical insights into selecting suitable clustering algorithms and dimensionality reduction techniques for high-dimensional datasets. It highlights the effectiveness of K-Means for data analysis tasks, contributing to improved clustering outcomes in similar applications