Filtering Reviews by Random individual Error

Abstract

Opinion mining from physician rating websites depends on the quality of the extracted information. Sometimes reviews are usererror prone and the assigned stars or grades contradict the associated content. We therefore aim at detecting random individual error within reviews. Such errors comprise the disagreement in polarity of review texts and the respective ratings. The challenges that thereby arise are (1) the content and sentiment analysis of the review texts and (2) the removal of the random individual errors contained therein. To solve these tasks, we assign polarities to automatically recognized opinion phrases in reviews and then check for divergence in rating and text polarity. The novelty of our approach is that we improve user-generated data quality by excluding error-prone reviews on German physician websites from average ratings.