PROJECT TITLE :
Robust Ensemble Clustering Using Probability Trajectories
Although many successful ensemble clustering approaches are developed in recent years, there are still two limitations to most of the present approaches. 1st, they mostly overlook the issue of unsure links, that could mislead the consensus method. Second, they typically lack the power to include international information to refine the local links. To deal with these 2 limitations, in this paper, we propose a unique ensemble clustering approach based on sparse graph illustration and chance trajectory analysis. In particular, we gift the elite neighbor choice strategy to spot the uncertain links by domestically adaptive thresholds and build a sparse graph with a little range of in all probability reliable links. We argue that a tiny range of in all probability reliable links will cause significantly higher consensus results than using all graph links no matter their reliability. The random walk process driven by a brand new transition likelihood matrix is utilized to explore the world info within the graph. We derive a unique and dense similarity measure from the sparse graph by analyzing the likelihood trajectories of the random walkers, based mostly on which two consensus functions are additional proposed. Experimental results on multiple real-world datasets demonstrate the effectiveness and efficiency of our approach.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here