KETEPATAN KLASIFIKASI PEMILIHAN METODE KONTRASEPSI DI KOTA SEMARANG MENGGUNAKAN BOOSTSTRAP AGGREGATTING REGRESI LOGISTIK MULTINOMIAL

Classification is one of the statistical methods in grouping the data compiled systematically. Classification problem rises when there are a number of measures that consists of one or several categories that can not be identified directly but must use a measure. classification methods commonly used in studies to analyze a problem or event is logistic regression analysis. However, this classification method provides unstable parameter estimation. So to obtain a stable parameter multinomial logistic regression model used bootstrap approach that is bootstrap aggregating (bagging). The purpose of this study was to compare the accuracy of the classification multinomial logistic regression models and bootstrap aggragatting model using the data of family planning in Semarang. From the results of bagging multinomial logistic regression obtained classification accuracy in replication bootstrap most 50 times at 51%, this model is able to decrease the classification error of up to 2% compared to the multinomial logistic regression model with a classification accuracy of 49%. Keywords : logistic regression, bootstrap aggregating, accuracy of classificatio