ABSTRACT
Breast Cancer (BC) is one of the most prevalent forms of Cancer among women. Premature diagnosis of BC is crucial to the survival of the patient. Here we implement an algorithm designed to diagnose and forecast breast cancer using a multi-layer perceptron (MLP) back-propagation technique that will help doctors diagnose the disease (benign, malignant). The proposed MLP includes an input layer, and, has inputs linked to the ten attributes of the data set. It has a hidden layer with five nodes (neurons). It leads to the pair outcomes: benign and malignant. The objective of our projected algorithm is to diagnose and classify the disease. MLP can help timely recognition of the cancer, and, therefore, can help to go for proper medication at early stage of cancerous development. This approach is tested on the (WBC) Wisconsin Breast Cancer dataset, resulted in 98.9 percent accuracy of classification using MLP back propagation.
. Key words: Neural-Network, Tumour, Prediction, Features, Training, Analysis, Multi-layer Perceptron
۱. INTRODUCTION
As per the reports, one of each eight women in the United States build ups breast cancer in their life span. It is one of the most critical diseases among women leading to their death. Premature identification needs an exact and consistent diagnosis system that allows physician to differentiate benign breast tumours from malignant ones without departing for surgical biopsy. Marcano-Cedeno et al. (2011) Breast cancer originates with an unrestrained partition of one cell and consequences in an observable mass called tumour [1]. The tumour can be benign or malignant. Katsis et at (2013) Breast Cancer causes several of the risk factors such as genetic, obesity, family history, having a first child after the age thirty,
not having children, aging, menstrual periods, not having children, smoking, drinking, that raise a women possibility of developing breast cancer [2]. Anil Arora et al. (2016) the precise diagnosis of the breast cancer is one of the critical problems in the medical field [3]. NN based MLP back propagation technique seems to be an efficient method for classification of breast cancer. This approach is pedestal on the WBC (Wisconsin Breast Cancer) and the taxonomy of dissimilar category of breast cancer datasets. The MLP back propagation is used to reduce the error rate of breast cancer and increase accuracy.
