Text mining wikipedia to discover alternative destinations

Abstract

This paper discusses an application of some statistical Natural Language Processing algorithms to a set of articles from Wikipedia about top tourist destinations. The objective is to automatically identify the key features of each destination and then discover other destinations which share similar sets of features. Through this a method is demonstrated by which meta data about each article can be extracted from the unstructured text and then used to answer complex discovery type queries. The paper compares an approach to automatically clustering similar destinations with a more user driven feature focused technique.