Wised Semi-Supervised Cluster Ensemble Selection: ANew Framework for Selecting and Combing MultiplePartitions Based on Prior knowledge

The Wisdom of Crowds, an innovative theory described in social science, claimsthat the aggregate decisions made by a group will often be better than those of itsindividual members if the four fundamental criteria of this theory are satisfied. Thistheory used for in clustering problems. Previous researches showed that this theorycan significantly increase the stability and performance of learning problems. As asolution, this paper proposes a new methodology of using WOC theory forevaluating and selecting basic result partitions in semi-supervised clusteringproblems. This paper introduces new technique for reducing the data dimensionsbased on supervision information, a new semi-supervised clustering algorithm basedon k-means for generating basic results, a new strategy for evaluating and selectingbasic results based on feedback mechanism, a new metric for evaluating diversity ofbasic results. The results demonstrate the efficiency of proposed method s aggregatedecision-making compared to other algorithms.

Wised Semi-Supervised Cluster Ensemble Selection: ANew Framework for Selecting and Combing MultiplePartitions Based on Prior knowledge Keywords:

Semi-Supervised Learning , Cluster Ensemble Selection , Wisdom of Crowds , PairwiseConstraints , Constraint Projection