Improve Performances and Efficiency in Clustering by using Single Pass Seed Selection Algorithm

We widely use k-means method for clustering technique for various applications. However, the k -means often converges to nearby superior and the end result depends on the initial seeds. Inappropriate desire of initial seeds may also yield terrible outcomes. K means++ is a manner of initializing k-means by way of choosing preliminary seeds with particular possibilities. Due to the random selection of first seed and the minimal likely distance, the k-means++ additionally outcomes different clusters in special runs in extraordinary range of iterations. In this examine we proposed a technique referred to as Single Pass Seed Selection (SPSS) algorithm as modification to k-means++ to initialize first seed and probable distance for k-means++ based at the point which became near extra wide variety of other factors within the statistics set. We evaluated its overall performance by making use of on various datasets and compare with k-means++. The SPSS set of rules turned into a single skip set of rules yielding particular solution in much less variety of iterations whilst in comparison to ok-means++. Experimental effects on real information sets from UCI established the effectiveness of the SPSS in generating regular clustering effects. Conclusion: SPSS performed well on high dimensional facts units. Its performance improved with features in the data set; specifically while range of capabilities we recommended the proposed method.