The breast cancer is a life-threatening disease observed among females all over the world. Detection and analysis of the disease is a significant part of data mining research. Classification as an essential data mining procedure also helps in clinical diagnosis and analysis of this disease. In our study, we proposed a novel Neuro-fuzzy classification based method. We applied our method to three benchmark data sets from the UCI machine learning repository for detection of breast cancer; they were namely Wisconsin Breast Cancer (WBC), Wisconsin Diagnostic Breast Cancer (WDBC), and Mammographic Mass (MM) data sets. Our objective was to diagnose and analyze breast cancer disease with the ...view middle of the document...
A medical survey in 2008 reveals that throughout the world breast cancer occurs in 22.9% of all cancers in women and it also causes 13.7% of cancer deaths in them . Medical prognosis and survival rates for the disease mostly depend on the type and stage of cancer, method of treatment, and geographical location of the patient.
The diagnosis of breast cancer normally begins when a patient or the physician finds a mass or abnormal chemical change on a mammogram by screening; or an irregularity in the shape of a woman’s breast by clinical test or self-examination. Generally the doctors study these factors when choosing a diagnostic test— medical condition and age, type of cancer believed likely, severity of indications, and previous clinical examination results. Physicians use diagnostic analysis for cancer and determine if it has spread to other parts of the body outside the chest area. For most cancer types, biopsy is the only way to diagnose cancer properly. If a biopsy is unacceptable, the clinician may suggest other medical tests that will help establish a diagnosis. Therefore, the well-known imaging tests like diagnostic using mammography, MRI or ultrasound might be done to determine the severity of breast cancer.
Classification as an important data mining technique can be used to diagnose and analyze breast cancer so that the disease could be detected at an early stage. The earlier detection of the disease followed by proper medical treatment might save the life of the patient from danger. That is why such analysis is very much important in the field of medical science as well as in Bioinformatics.
Artificial Neural Network (ANN) [5, 6, 7] is a popular data modeling tool that can perform intelligent tasks similar to the human brain. It is well-known for high precision and high learning ability even when a very little information is available. One of the reliable and efficient methods of data classification from the ANN domain is the Multilayer Perceptron (MLP) algorithm [8, 9]. A typical MLP network model has a single input layer, at least one hidden layer, and a single output layer. The nodes in the input layer and in the hidden layer(s) are associated with weights. A MLP model exhibits two modes of operation: feedforward and backpropagation.
Support Vector Machine (SVM)  is a very powerful supervised learning model used in the machine learning field. It constructs one, or more than one, hyper-plane in a high-dimensional feature space for classification, regression or other analysis tasks. SVM employs a hyper-plane to differentiate between classes. When classes overlap, a hyper-plane is selected to minimize the error of data points along or across the boundary line between classes; these points are referred to as the support points or support vectors.
Due to the presence of ambiguity in input data, overlapping boundaries among classes, and indefiniteness in defining features some uncertainties can still arise at any stage of...