TEXT MINING UNTUK KLASIFIKASI PENGADUAN PADA SISTEM LAPOR MENGGUNAKAN METODE C4.5 BERBASIS FORWARD SELECTION

Ali Sofyan
Stefanus Santosa

Jurnal Teknologi Informasi Cyberku
Vol 12 No 1 (2016): Jurnal Teknologi Informasi CyberKU Vol. 12, no 1

Report of public complaints on the site becomes a medium of communication between the community and government agencies. The number of incoming documents every day be a source of information to measure the services of government agencies. Classification of documents is very important to do otherthan to ensure that the intended objectives of the institution, as well as to classify complaints fit the category. C4.5 algorithm is one of the algorithms that can be used for classification. There were some complaints classification research. This study aims to apply the classification of complaints by algritma C4.5 with a selection of features to improve the accuracy of classification. Results of experiments with methods of research division of the number of datasets, cross validation, classification with and without features. The test results obtained by testing the value of the best accuracy with 550 documents with forward selection, with cross valiadtion 9folds with a value of 85.27%. precission 87.8% and 85.3% recall