Analyzing Twitter Data with Preferences.

Today Twitter is one of the most important sources for information distribution. But finding useful and interesting tweets on a specific topic is a non-trivial task, because there are thousands of new posts every minute. In this paper, we describe our preference-based search approach on Twitter messages, which allows users to get the best possible results. For this, we introduce a new CONTAINS preference constructor to search on full-text data, use NLP techniques to handle natural language mistakes, and present experiments.