EJONS INTERNATIONAL JOURNAL ON MATHEMATICS, ENGINEERING & NATURAL SCIENCES ISSN 2602 - 4136

Quick Access


Bu Dergi DOI ve Crosscheck üyesidir


Abstract



Clustering is the decomposition of similar objects in the data sequence to take place in the same groups. Although kmeans from partitioned cluster analysis algorithms is popular in clustering, it has three fundamental deficiencies in terms of insufficient computation, user-defined number of clusters and local minimum in search. The X-means algorithm is an alternative algorithm that, after every run of the k-means, can find the number k in the specified range, giving local decisions about which subclusters of the existing centers should be better fit to fit the given value. A new algorithm has been proposed to optimize the Bayesian Informatization (BIC) with X-means and to efficiently search the number of clusters and clusters. In this study, k-means and x-means clustering were applied to the same dataset with the help of WEKA. It is seen that the work done with K-means is tested with Silhoutte coefficient, x-means with BIC, and x-means clustering algorithm produces more effective results than k-means clustering algorithm.



Keywords
Clustering , k-means, x-means



References

Advanced Search


Announcements

    Mart Sayısı

    VERBİS'E KAYIT S�RELERİ HAKKINDA DUYURU | Kırklareli Ticaret ve ...

    Mart ayında yayınlanacak Sayımız için  25 Şubat Tarihine kadar çalişmanızı Yükleyebilirsiniz



Address :Kazakh National Women's Pedagogical University, Almatı, Kazakhistan
Telephone :+7 778 921 0336 - 0 534 349 92 73 Fax :
Email :ejonsjournal@gmail.com

Web Yazılım & Programlama Han Yazılım Bilişim Hizmetleri