Effect Of The Learning Rate, Hidden Layer, And Epoch In Lung Cancer Prediction

Abstract

Lung cancer is one of the deadliest diseases in the world after breast cancer, with a prevalence rate of 11.4%. This research aims to predict the level of accuracy of whether a person is diagnosed with suspected cancer or not. The method used in this study uses a backpropagation neural network using a combination of the learning rate, hidden layer, and epoch values. The results showed that the learning rate value of 0.1, the value of hidden layer 100, and epoch 1000 resulted in predictions with the best accuracy value of 97%.