Ensembled Heuristic Iterative Expected Maximization with BrownBoost Data Clustering for Uncertain Data Mining

摘要

Clustering on uncertain data needs to be handled in order to generate significant knowledge patterns. Clustering is the data mining tasks to group similar information or data. Though the conventional algorithm groups similar data for mining the uncertain data, minimizing the error rate and clustering time and improving the accuracy turn out to be the major issues in clustering of uncertain data. In this paper, Ensembled Heuristic Iterative Expected Maximization with BrownBoost Data Clustering (EHIEM-BBDC) technique is proposed in which the number of base learners are constructed using Iterative Expected Maximization and the heuristic is applied to speed up the process of clustering. The BrownBoost technique is applied for improving the clustering accuracy by combining the base learners to form strong cluster. Experimental evaluation of proposed EHIEM-BBDC technique and existing methods are carried out with the El Nino dataset taken from the UCI machine learning repository. The results have shown that the proposed technique outperforms well to mine the uncertain data through the high clustering accuracy with minimum time as well as less false positive rate.