Artificial Intelligent Models for Breast Cancer Early Detection
Erwin Halim, Pauline Phoebe Halim, Marylise Hebrard · 2018
Cancer prevalence increases year by year and the medical management develops along with it. As the number increases, breast cancer which contributes most to the number become major attention. Breast cancer detection can be done using many methods but their main purpose are all same. Considering the current situation, Machine Learning (ML) application to support cancer detection have the ability to provide better diagnostic result as in 97.40/0 of the cases without ML participation, malignancy was to be found on surgery while 30.6% surgeries conducted on benign lesion can be prevented. This paper propose an early detection model for breast cancer which will combine models from previous researches to become functional for many detection methods. Methods that had been proven to show accuracy in their specific field and suggested to be used in parallel: DWT-based multi-resolution MRF (MMRF) segmentation for mammography, MLP for histologic examination, and k-NN - SVMRFE method for gene identification which will be effective by learning from experience of at least 450 datasets of positive breast cancer result for each method and refers to Wisconsin Breast Cancer Diagnosis (WBDC). The research itself will be conducted from October 2018 until March 2019 and the final result will be applied to conduct further research in ML application in breast cancer early detection.