BREAST CANCER ANALYSIS USING LOGISTIC REGRESSION

Abstract

In this study, the diagnosis of breast cancer from mammograms is complemented by using logistic regression. The radiologists can use the results to make a proper judgment as to the presence of breast cancer. Data were obtained from survey questions completed by the radiologist during his observation of the patients. The results using logistic regression cross tabulation was to obtain the significant values between the breast cancer factors. The classification table from 130 samples shows the occurrence from prediction and observation samples, producing percentage of correct classification for mammogram results is 91.5%. The accuracy is compared with validated samples which are 46 samples and the percentage of correct classification is 67.4%. The analysis for mammograms screening using parameter estimation is to identify all the factors that were available in the survey. The presence of mass, architectural distortion, skin thickening, and calcification had high odds of getting breast cancer.