Flow based clustering algorithm for tourism search engine

This paper introduces a flow based clustering algorithm for tourism search engine. Unlike the general tourism search engines such as www.tripadvisor.com, www.qunar.com and www.kayak.com etc. to return the users´ queries huge amount of web page links, this algorithm helps the tourism search engine create a list of words which serve as suggestions to expand and update the users´ queries. It is much different from the previous clustering algorithms which cluster the results by the similar subjects. This algorithm clusters the search results by the inside vector distance representing each relationship between the web pages. In this manner, these representatives belong to different clusters merging into something like different flows. The experimental results indicate that the algorithm performs an unexpected usefulness for the users´ queries. Meanwhile, through the comparison with some often used algorithms and an online survey, it shows this algorithm has an accepted performance.