KLASIFIKASI SENTIMEN PEMBELI BERDASARKAN LAYANAN SMS “SUARA KONSUMEN” TERHADAP PRODUK MENGGUNAKAN METODE K-NN

David Addiwijaya Hanz (Universitas Kristen Duta Wacana)
Budi Susanto (Universitas Kristen Duta Wacana)
Lukas Chrisantyo (Universitas Kristen Duta Wacana)

Jurnal Informatika
Vol 11, No 1 (2015): Jurnal Teknologi Komputer dan Informatika

Limited number of characters in an SMS causes the use of words and the structure of solid and compact but obviously become very important in presenting the content and purpose of the sender of the SMS. This research studied how the use of SMS classify structures and limited words to get the sentiment of an SMS service "Suara Konsumen" through text mining approach. This study uses K-NN algorithm in the classification, besides the K-NN algorithm of this study will be weighted words using TF-IDF methods, Feature Selection in the selection of words also Cosine Similiarity in measuring the degree of proximity between documents. In determining the success, measured the accuracy of the documents the trial and the results are very accurate with an average value of 89% accuracy on the variation of K and Feature Selection.