Pembuatan Aplikasi Klasifikasi Otomatis Laporan Keluhan Warga di Polrestabes Semarang

Abstract

This research aims to simplify the classification document of report complaints in Polrestabes Semarang. This research used cosine similarity algorithm and nearest neighbor algorithm on classification process. Cosine similarity measure between two vector of an inner product space. Nearest neighbor calculate vector distance of the document to another. The result show that the accuracy of cosine similarity algorithm is better than nearest neighbor algorithm. It has the value 93,334% for cosine similarity algorithm and 80% for nearest neighbor algorithm.