Extraction of key topics from online text reviews

Abstract Though it has been a subject of active research for a while, extraction of key phrases from unstructured textual data is not a completely mature technology. Apart from the usual challenges in computational linguistics such as synonymy and polysemy, there may be additional domain specific challenges. So, a state-of-art in one domain may not be so in a different domain and there is hardly any universally acceptable and completely accurate solution. In this paper, we evaluate and compare different approaches for key topic extraction from unstructured textual data found in online review and rating portals.