Comparative analysis of Malignancy prediction of Breast Cancer cells using Logistic Regression& K Means Algorithm

Abstract

Cancer is treated as one of the major health concerns of the country and it is caused when some abnormality happens to a cell. Among cancer, one of the most common cancer that affects the female gender is breast cancer and there are many new cancer cases as well as death reported globally. So, an early diagnosis of cancer can provide clinical suggestions to recover or improve the survival rate significantly. This paper suggests a comparative analysis of the detection of breast cancer using one of the machine learning algorithms like logistic regression and K Means Algorithm. An open dataset provided by the University of Wisconsin Hospital at Madison, Wisconsin, USA will be used to implement the algorithm. There are 569 instances of data available in the open dataset with its classification as malignant and benign. The datasets undergo preprocessing followed by the implementation of algorithms from which results are visualized using the orange tool. Models are trained and tested from which results are displayed in different forms. The classification is depicted using a confusion matrix as well as through a ROC curve while MDS and Silhouette plot is used for K-means clustering. Comparison between logistic regression and k means clustering is observed along with a comparison of different literature reviews with research in hand. It was concluded that logistic regression is a better predictive model based on the given requirements and usage of different machine learning tools significantly affects the results and accuracy of the model.