Exploiting Silhouette Principle Component For Dimension Reduction In Breast Ultrasound Images Classification

Kartikadarma, Etika
Fanani, Ahmad Zainul
Pujiono, Pujiono
Affandy, Affandy
Wulandari, Sari Ayu

International Journal of Artificial Intelligence Research
Vol 8, No 1 (2024): June 2024

This paper proposes the use of the Dimensional Reduction method with the Silhouette Principle Component (SPC) Approach to improve the classification of breast ultrasound images. The PCA method is used to reduce the dimensions of medical images to improve classification, with MobileNet-v2 and DenseNet-121 as the optimal classification algorithm choices. The results show that the SPC method succeeded in producing efficient feature representation with data sizes that are almost the same as the original data, while PCA produces greater dimensionality reduction. The SPC model also shows the best performance in terms of accuracy and loss. This research makes a significant contribution to the development of more sophisticated and efficient medical image analysis techniques to support the diagnosis and treatment of breast cancer.