Best feature selection and data visualization for breast cancer prediction

Abstract

Breast cancer (BC) is one of the most common cancers among women worldwide, representing the majority of new cancer cases and cancer-related deaths according to global statistics, making it a signifcant public health problem in today's society. In medical diagnosis, the forecast of an infection goes about as a signifcant center in breaking down the therapeutic pictures. The undesirable cell development in any piece of the organ is known as tumor. The tumor might be favorable or harmful. Threatening tumor is viewed as the most risky tissue. There are diferent specialists learned about the forecast of bosom malignancy. This paper aims to review on various data set techniques that are specifcally considered on breast cancer prediction and also to investigate which feature set is responsible for the disease and rapid growth of cancer cells as we are selecting the best features. From primarily given data set we can measure which parameter is responsible for cancer cells and which features can make the nearest perfect outcomes. It is presently possible to make precise Computer Aided Diagnosis (CAD)[7] framework so as to make the whole procedure of distinguishing a dangerous tumor more asset profcient and efcient through appropriate usage. This paper displays the relative investigation of various machine learning calculations and their outcomes in anticipating destructive tumors. For example, Decision Tree, Support Vector Machine, K-Nearest Neighbors, Linear Discriminant Analysis, Naive Bayes and Logistic Regression with and without PCA on a dataset with 30 highlights removed from a digitized picture of a Fine Needle Aspirate (FNA)[19] of a breast mass. Profound learning models like Artifcial Neural System and Convolutional Neural Network are utilized and their exhibitions are looked at.