Breast and Thyroid Cancers Classification using Scaled Conjugation Backpropagation, Levenberg-Marquardt, and Bayesian Regularization
Muhammad Albik Ghalela, Yessi Jusman · 2023
This neural network research is aiming to classify breast cancer and thyroid. The methods used in these experiments are scaled conjugation backpropagation, Levenberg-marquardt, and Bayesian regularization within 10 cross-validation training. The average accuracy of breast cancer classification is 96,92% with scaled conjugation backpropagation, 97,08% with Levenberg-marquadt, and 99,59% for Bayesian regularization. While average accuracy of thyroid classification with 94,18% with scaled conjugation backpropagation, 98,78% with Levenberg-marquardt, and 99,59% with Bayesian regularization. Finally, the result of this study is the differential of each neural network algorithm that can be useful for image medical classification in both breast cancer and thyroid.