Analysis of Decision Tree and K-Nearest Neighbor Algorithm in the Classification of Breast Cancer.

Harikumar Rajaguru,Sannasi Chakravarthy S R

doi:10.31557/apjcp.2019.20.12.3777

Harikumar Rajaguru, Sannasi Chakravarthy S R

Open Access

https://doi.org/10.31557/apjcp.2019.20.12.3777

Copy DOI

Abstract

Objective:The death rate of breast tumour is falling as there is progress in its research area. However, it is the most common disease among women. It is a great challenge in designing a machine learning model to evaluate the performance of the classification of breast tumour. Methods:Implementing an efficient classification methodology will support in resolving the complications in analyzing breast cancer. This proposed model employs two machine learning (ML) algorithms for the categorization of breast tumour; Decision Tree and K-Nearest Neighbour (KNN) algorithm is used for the breast tumour classification. Result:This classification includes the two levels of disease as benign or malignant. These two machine learning algorithms are verified using the Wisconsin Diagnostic Breast Cancer (WDBC) dataset after feature selection using Principal Component Analysis (PCA). The comparison of these two ML algorithms is done using the standard performance metrics. Conclusion:The comparative analysis results indicate that the KNN classifier outperforms the result of the decision-tree classifier in the breast cancer classification.

Highlights

Breast cancer refers to a disease which is found more among the women in developed and in developing nations
This proposed model employs two machine learning (ML) algorithms for the categorization of breast tumour; Decision Tree and K-Nearest Neighbour (KNN) algorithm is used for the breast tumour classification
The comparative analysis results indicate that the KNN classifier outperforms the result of the decision-tree classifier in the breast cancer classification

Summary

Introduction

Breast cancer refers to a disease which is found more among the women in developed and in developing nations. Regular screening of breast is required by the doctors for the earlier discovery of the disease. This earlier discovery of breast cancer denotes the identification of disease before the symptoms are felt by the patients This denotes the observation of tissues in the breast-part of the patients for any abnormal lumps. At this stage, a fine-needle aspiration (FNA) biopsy procedure is needed if any lump or mass is appeared on the breast while screening (Sannasi Chakravarthy et al, 2019). This biopsy method is used to collect a few cell samples in the region of lump occurrence in breast It is a simple procedure and it does not require any hospitalization for the patients. These features are utilized by the Decision-Tree algorithm and KNN algorithm to perform

Materials and Methods

Results

Methodology

Discussion