PENERAPAN ALGORITMA K-NEAREST NEIGHBOR DALAM KLASIFIKASI JUDUL BERITA HOAX

With the rapid development of information technology, especially in Indonesia, is more easily obtained through online media. Therefore, dissemination media becomes uncontrollable and a lot not accordance with facts or can be said to hoax. Readers should careful when reading news headlines avoid hoaxes. The purpose this research find out how apply K-Nearest Neighbor (KNN) algorithm classifying including hoaxes In process, classification non-hoaxes uses KDD method text mining goes several stages, namely preprocessing, word weighting TF-IDF using KNN algorithm. There are 3 scenarios data split 90:10, 80:20, 70:30. Evaluation done by confusion matrix. results study highest accuracy 93.33% k value 90:10 scenario. So, suitable for hoax titles.