Comparitive study on the performance of various clustering approaches

Karanjit Singh Tiwana, Prof. Saleema j. S.

Abstract


Clustering is a technique of discovering and grouping similar objects into a single logical unit. Clusters can be created based on similarities or dissimilarities between objects. Usually clusters employ a distance measure to calculate the similarity between objects. So, objects within a cluster are more similar to each other than objects in another cluster. This paper uses real-life datasets to compare the accuracy of the most popular algorithm in each approach.


Keywords:clustering; k-means; hierarchical based; density based; partition based; agnes; dbscan; distance measure

Full Text:

PDF


DOI: https://doi.org/10.26483/ijarcs.v8i3.3040

Refbacks

  • There are currently no refbacks.




Copyright (c) 2017 International Journal of Advanced Research in Computer Science