Klasifikasi Teks Pengaduan Pada Sambat Online Menggunakan Metode N-Gram dan Neighbor Weighted K-Nearest Neighbor (NW-KNN)

Annisya Aprilia Prasanti (Fakultas Ilmu Komputer, Universitas Brawijaya)
Mochammad Ali Fauzi (Fakultas Ilmu Komputer, Universitas Brawijaya)
Muhammad Tanzil Furqon (Fakultas Ilmu Komputer, Universitas Brawijaya)

Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Vol 2 No 2 (2018): Februari 2018

SAMBAT Online is a concrete application of E-Government in a web-based platform for complaints provided by Dinas Komunikasi dan Informatika Kota Malang (Diskominfo Malang). An incoming complaint text will be categorized into various areas of the SKPD. With that being said, in order to make the job of the super admin easier in organizing and determining an SKPD category, as well to organize a complaint text and improve the time efficacy, a method of text classification is paramount. NW-KNN is an upgraded algorithm of the traditional KNN algorithm. Generally, the closest neighboring distance calculations will use Cosine Similarity with bag of words for feature extraction. Bag of words is a feature extraction that ignores the order of words of a sentence altogether. To improve the algorithm despite the deficiency, this research will use supporting method for feature extraction, which is called as N-Gram. The result in this research indicated that NW-KNN with neighboring value k = 3 and N-Gram with Unigram have the highest f-measure's value with 75.25%.