The development of information every day is increasing. Public complaint is one form of information on the Internet is growing each day according to the number of people who make a complaint. In the management of complaints frequent errors in aduannya groupings so as to make the admin must work longer to perform grouping or classification of complaints. Such information becomes a media which is used for data mining research. One of the functions of data mining is classification. Naïve Bayes is one of the methods used for classification, one for the classification of documents or text. The classification is very useful for grouping data or documents by category. This will simplify the user data or documents in the search process. This research was conducted by applying the method Naïve Bayes for classification societies complaint data and algorithms Information Gain for the selection of attributes in order to improve the accuracy of the classification of public complaints. The test results by using 150 training data and testing the data 60 Naïve Bayes algorithm using attribute selection results without accuracy is 63.33%. whereas on testing Naïve Bayes algorithm using Information Gain attribute selection with the same data results are increasing even with k = 5. The best accuracy results found in this study was 86.67% using the selection attribute by 55.
Copyrights © 2019