Breast Cancer Auxiliary Diagnosis Combining Case- Based Reasoning And Classifiers

This study employs Na?ve Bayes,K Nearest Neighbor(K-NN)and Case Base Reasoning(CBR)approach to assemble a Breast Cancer(BC)diagnosis model with the purpose of optimizing the retrieval process of CBR by raising its accuracy rate.In light of the counsel and help of health specialists we selected a sample data from of the Mozambique Breast cancer Data set from the Central Hospital of Maputo(HCM)in Mozambique within the years 2014 to 2016.A number of about 1200 patients were chosen as cases in the data set for breast cancer diagnosis,from where the training and testing sets was generated.This paper hence proposes an intelligent model for the diagnosis of breast cancer which combines Na?ve Bayes,CBR and KNN.The main steps in implementing the model include:(1)adopting Na?ve Bayes model to classify the set into two distinct classes(2)employing K-NN algorithm into CBR for the retrieval of most similar cases.In the first phase Na?ve Bayes was utilized to make estimation on whether a patient has a Malign or a Benign tumor,comparisons with K-NN and J48 Decision Tree classifiers were made and Na?ve Bayes showed an outstanding performance with accuracy rate of 95%.In the second phase we tested the selected value of k,the results show the accuracy of 99%.And the retrieval results of the suggested framework for diagnosis after implementing K-NN showed cases with high similarity rates among retrieved cases with a minimum distance all the way down to 0.13.The results show that the implemented model is capable of integrating Na?ve Bayes and CBR for Breast Cancer diagnosis.It can supply a supporting system for health practitioners in the diagnosis of breast cancer,allowing lessening of diagnosis inaccuracy,as well as enhance the quality of healthcare Assessment.