A game with purpose to filter spams from Indonesian Twitter trending topics

Trending topic is a place where topic that matters in certain region up to the whole world placed on Twitter. Besides directly mentioning other users or using direct messages, riding on a trending topic that Matters in Twitter is another common practice to spam. The ease of use of this technique makes it is very easy to spammers who want to get their messages exposed. On the other side, current condition on Twitter trending topic suggests that Twitter might be overwhelmed by spammers. It relies on their massive user but tend to hide spam reporting option, make spam reporting a secondary activity. In this paper, we propose a content based method to create spam filter in Indonesian Twitter trending topics. This is done by asking several people to play the game. Inside the game, there is a tweet determination process, a proses which gives people who play the game a pair of hashtag that is usually marks a trending topics and tweet corresponding with it resulted in possible spam words statistics. This statistics used as a spam filter was tested using 648 tweets that was fetched from Twitter. The filter checked every words of a tweets inside the database. This database also filled with statistic that was created by people who plays the game. This Method was able to detect spam tweets with 98 percent accuracy.