The Modified Perceptron Algorithm for Breast Cancer Diagnosis
Amal M. El Nawasany, Ahmed Fouad, M. El-Sayed Waheed · CiiT international journal of data mining and knowledge engineering · 2013
Breast cancer is the most common cancer in women worldwide. It is also the principle cause of death from cancer among women globally. Early detection of this disease can greatly enhance the chances of long-term survival of breast cancer victims. Classification of breast cancer as malignant or benign can be achieved using many techniques. Artificial neural network (ANN) is one of these techniques. This paper presents a new ANN algorithm named as modified perceptron algorithm for data classification (MPADC) has been formed to classify a nonlinearly separable data. A new termination criterion is added to the original perceptron algorithm by using evaluation vector in order to monitor the accuracy that helps in setting predefine termination accuracy interval to avoid endless execution times as if the classification problems are not linearly separable then percerptron will be executed infinite number of times. In order to detect the classification accuracy of the proposed algorithm, we apply it to the Wisconsin breast cancer original database (WBCD) data. The average classification accuracy of the developed MPADC algorithm is 99.54%. Among the studies use neural network classification algorithms for the same breast cancer database, our algorithm appears to be very promising.The results suggest that MPADC is faster, more feasible and more accurate classification algorithm than the results obtained from related previous algorithms.