Breast Cancer Diagnosis Using Three Supervised Learning Algorithms

Abstract

Vrast Cancer is one of the most dangerous forms of illness. Almost 12,000 cases of Breast Cancer end in death anually in the UK. In order to help with this case, previous research has been done to discover early the cancer type: benign or malignant. Artificial Neural Networks(ANN), K-Nearest Neighbor(KNN) and Decision Tree(DTree) are three kinds of supervised learning algorithms each of which has different ways to classify data. What makes this research a challenge is to compare the accuracy values in Breast Cancer prediction. The Breast Cancer Wisconsin data contains three kinds of data: mean, standard error, and largest which were taken from 569 patients. This adds thecomplexity of way each algorithm performs. The result shows that ANN and KNN have better performance than DTree. The study also offers recommendation for doctors regarding which kind of data should be considered first in Breast Cancer diagnosis. In comparison to the previous research, this study finding has better accuracy.

 

Keywords: Breast Cancer Wisconsin, Artificial Neural Network, K-Nearest Neighbor, Decision Tree.