IJCTT-International Journal of Computer Trends and Technology
The clustering algorithm, so far, are inefficient in finding interested patterns as they reckon on the position of data items which results in large number of illegitimate patterns. The generated illegitimate patterns have low association between them providing knowledge which is of nominal significance to the user. The existing algorithm results in poor propensity among the data items. To perform the exception of escalated patterns, the proposed algorithm uses a nominal threshold value as decided by the user. The escalated patterns will be excerpted using e-confidence and mining for knowledge will be applied on these patterns.