In this era, a rapid thriving Internet occasionally complicates users to retrieve news category furthermore if there are plentiful of news to be categorized. News categorization is a technique can be used to retrieve a category of news which gives easiness for users. Internet has vast amounts of information especially at news. Therefore, accurate and speedy access is becoming ever more difficult. This paper compares a news categorization using k-Nearest Neighbor, Naive Bayes and Support Vector Machine. Using vary of variables and through a several steps of preprocessing which proving k-Nearest Neighbor is producing a capable accuracy competes with Support Vector Machine whereas Naive Bayes producing just an average result, not as good as k-Nearest Neighbor and Support Vector Machine yet as bad as k-Nearest Neighbor and Support Vector Machine ever reach. As the results, k-Nearest Neighbor using correlation measurement type produces the best result of this experiment. Abstrak Pada zaman sekarang, perkembangan internet yang begitu pesat kadang-kadang mempersulit pengkategorian berita apalagi jika ada banyak berita yang harus dikategorikan. Kategorisasi berita adalah teknik yang bisa digunakan untuk mengambil kategori berita yang memberi kemudahan bagi pembaca berita. Internet memiliki sejumlah besar informasi terutama di berita. Oleh karena itu akses yang akurat dan cepat menjadi semakin sulit. Makalah ini mengulas kategorisasi berita dengan menggunakan k-Nearest Neighbor, Naive Bayes dan Support Vector Machine. Penggunaan berbagai variabel dan melalui beberapa tahap preprocessing yang membuktikan k-Nearest Neighbor menghasilkan akurasi yang mampu bersaing dengan Support Vector Machine sedangkan Naive Bayes hanya menghasilkan hasil rata-rata, tidak sebagus k-Nearest Neighbor dan Support Vector Machine namun juga tidak memberikan hasil yang seburuk yang dihasilkan dengan teknik k-Nearest Neighbor dan Support Vector Machine. Sebagai hasil yang didapatkan, k-Nearest Neighbor dengan tipe pengukuran correlation memberikan hasil yang terbaik dan konsisten selama eksperimen dengan parameter dan validasi data yang berbedabeda.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.