Automatic clusters recognition method based on OPTICS reachability-plot
Yueming Lu · Journal of Computer Applications · 2012
In order to get better clustering performance for search results whose topic character is not obvious,a novel automatic clusters recognition method based on OPTICS(Ordering Points To Identify the Clustering Structure) reachability-plot,Smooth method,was proposed.First,OPTICS algorithm was used to get the reachability-plot of search results.Afterwards,the reachability-plot was smoothed by moving average method to remove burr.The clustering results were finally achieved by recognizing the dented regions according to peak values in the smoothed reachability-plot.Experimental results on data sets formed by search results show that the accuracy rate of Smooth method approximately achieved a 100% improvement than existing method.Smooth method combined with OPTICS algorithm can apply to Web search results clustering and make search results easier viewing.