A Bio-inspired Fuzzy Agent Clustering Algorithm for Search Engines

In general, web search engines respond to queries by returning a list of links to web pages that are considered relevant. However, these queries are often ambiguous or too general and the users end up browsing through a long list of items in order to find what they are actually looking for. And hence the idea to cluster web search results so that the output would be a list of labelled clusters. An algorithm based on the ASM (Ants Sleeping Model) is proposed. In the ASM model each data is represented by an agent, its environment being a two dimensional grid. The agents will group themselves into clusters by making simple moves according to some local environment information. At any step an agent can pro-actively decide to directly communicate with one of its fellows and choose to move accordingly, the moves being expressed by fuzzy IF-THEN rules. Thus the chance of getting trapped in a local optimum is minimized and hybridization with a classical clustering algorithm becomes needless. © Selection and peer-review under responsibility of FET11 conference organizers and published by Elsevier B.V.