Comparitive performance study of different clustering algorithms using different pre-processing techniques with different numbers of clusters on different evaluation parameters on Perfume Dataset
K-Mean Clustering works the best with data without preprocessing. It gave more balanced and efficient results. K-Mean Shift Clustering works equally good and best with data without preprocessing and after PCA. Also, it worked equally good even for different number of clusters. Hierarchical Clustering works best for normalised data. Overall K-means performed the best for Perfume Dataset.