Effect of noise on the performance of clustering techniques

The objective of the paper is to show the effect of noise on the performance of various clustering Techniques. Clustering is being widely used in many application including medical, finance and etc. Clustering may be applied on database using various approaches, based upon distance, density, hierarchy, and partition. The data item which is not relevant to data mining is called noise (e.g. out of bound and etc). Noise is major problem in cluster analysis, which degrades the performance of various clustering techniques in the term of time and space. Our purpose is to study how a particular clustering technique is responsive to the noise in the terms of time. We have used clustering techniques based upon the partitioned or non-hierarchical clustering, which forms the clusters in one step as opposed to several steps. We will introduce noise in the experimental database and the same will be used for clustering algorithm. Then the percentage of noise will be varied and the time required for clustering will be calculated. The observation so obtained will be used for analysis purpose. After that we will compare and analysis the performance of the clustering techniques based on the time required. Then we will conclude which algorithm is more realistic to noise.