Aplikasi Pendeteksi Opini Spam

Abstrak

Opinion mining aims to determine an opinion category, positive opinion, negative opinion, or neutral. Opinion mining implementation on public opinions, produces a percentage summary of positive opinions, negative opinions, or neutral from reader. In online political news context, this summary has advantage for readers to understand political situation which changed very fast and has advantage for online news manager to know respond from society towards presented news. Actually reader"Ÿs opinion also can be used to influenced public opinion, motivate appearance mislead spam opinion. Spam opinion must be detected in opinion mining, further spam op inion will be ignored.

This research purpose is opinion spam detection. This research used two methods, comments duplication checking and aberrant behavior detection. Duplication checking is done by comparing each existing comments on all articles by using the cosine similarity. This research can find some spam opinions by those methods.