IDENTIFIKASI FAKTOR PREDIKSI DIAGNOSIS TINGKAT KEGANASAN KANKER PAYUDARA METODE STEPWISE BINARY LOGISTIC REGRESSION

Abstract

The World Health Organization (WHO) reported that deaths caused by cancer in the world these last four years has increased significantly. The data also reflected in the increase in breast cancer cases. In Indonesia, two cases also the highest cases of adult female deaths. Based on Hospital Information System, the number of breast cancer patients either inpatient or outpatient care amounted to 28.7%. This fact revealed more than 40% of all cancers can be prevented with early detection cancer. Role of Information Technology can implemented by data mining techniques to shorten the diagnosing time, accuracy and selection of factors early detection of breast cancer. Stepwise binary logistic regression method has the advantage to add and subtract the independent variables in accordance with level of significance of the model. Based on the analysis of weighting method, the highest four variables that should be more aware is the area of cancer (area), fineness (smoothness), the number of dots (concave points) or the nucleus of cancer and grayish level of cancer (texture). So the accuracy and processing speed of diagnosis of the severity of breast cancer can be improved through this method.