AN ENHANCED K-MEAN CLUSTERING ALGORITHM

Sapna .

Abstract


K-Mean’s clustering algorithm is one of the most widely used partitioning algorithm used for grouping the elements. It is the fast, simple and can work with large datasets. But still it has some drawbacks like in the initial stage we have to tell the number of clusters. It can detect only spherical clusters. Number of iterations is more. Here we will propose an enhanced K-Means clustering algorithm which will basically work on the concept of partitioning dataset and reducing the number of iterations. It will abstract some features from two modified K-means algorithms. The benefit of partitioning is that we will be able to deal with larger datasets and the benefit of reducing iterations is that time taken for clusters formation will reduce and in this way the efficiency of the traditional K-means clustering algorithm is increased. The results of the proposed methodology, is applied on Enron dataset to find out spam emails in the spam email dataset.

Keywords


Data Mining, KDD, E-Mail, Spam, Ham, Spam Filter, K-Means Algorithm, K-Means Update Algorithm.

Full Text:

PDF


DOI: https://doi.org/10.26483/ijarcs.v8i9.5186

Refbacks

  • There are currently no refbacks.




Copyright (c) 2018 International Journal of Advanced Research in Computer Science