I.J. Information Technology and Computer Science, 2017, 3, 71-79 Published Online March 2017 in MECS (http://www.mecs-press.org/) DOI: 10.5815/ijitcs.2017.03.08 Copyright © 2017 MECS I.J. Information Technology and Computer Science, 2017, 3, 71-79 Priority Based New Approach for Correlation Clustering Aaditya Jain M.Tech Scholar, Department of Computer Science & Engg., R. N. Modi Engineering College, Rajasthan Technical University, Kota, Rajasthan, India E-mail: aadityajain58@gmail.com Dr. Suchita Tyagi Associate Professor, Department of Computer Science & Engg., Sushila Devi Bansal College of Technology, Indore, MP, India E-mail: suchitatyagi625@gmail.com Abstract—Emerging source of Information like social network, bibliographic data and interaction network of proteins have complex relation among data objects and need to be processed in different manner than traditional data analysis. Correlation clustering is one such new style of viewing data and analyzing it to detect patterns and clusters. Being a new field, it has lot of scope for research. This paper discusses a method to solve problem of chromatic correlation clustering where data objects as nodes of a graph are connected through color-labeled edges representing relations among objects. Purposed heuristic performs better than the previous works. Index Terms—Clustering Problems, Correlation Clustering, Chromatic Balls, and Priority Based Chromatic Balls. I. INTRODUCTION Clustering is an unsupervised from of machine learning aiming at grouping of data objects in a way that similar objects fall in the same group specifically called a “cluster”. The traditional clustering algorithms like k- means [1] and fuzzy c-means [2] use the notation of similarity or closeness among objects to group them. Thus, they view objects as having binary or fuzzy relationship between them. The binary relationship categorizes which clusters are similar and should be grouped in the same cluster using some similarity / distance metric between them. The fuzzy relations, on the other hand, deduce a percentage of similarity between data objects, with the ones with higher percentage probable to fall in the same cluster. In real world problem, the relations among objects are more complex. Like those existing among people in social networks, who have varying kind of relationships family, professional, friendly etc. Such scenarios of complex relations also exist in authored documents library, protein-protein interactions etc. Scenarios discussed above are best described through categorical relationships among objects, easily represented through graphs. Using graphic is advocated due to  They are flexible and intense data structures.  They can be easily ranged from very simple to very complicated relationships.  They can be used to represent many kinds of relations, whether independent or co-existing. Once a graph has been formed, the problem of analysis is converted into problem of partitioning the graph. Bansal et al. defined the problem of Correlation Clustering in [3]. It was successful enough to eradicate all the issues encountered in the traditional clustering algorithms so is being used in many applications like parallel and distributed system, pattern recognition, and image segmentation. Bonchi et al [4] further extended the concept of correlation clustering to chromatic correlation clustering by assigning colors to edges instead of positive or negative signed labels as used in correlation clustering. This paper presents a contribution in the direction of solving chromatic correlation clustering problem through revisiting the work of Bonchi et al [4, 5]. A Priority Based Chromatic Balls algorithm is presented to increase the probability of better solution of the algorithm and keeping its advantages of speed retained. The rest of the paper is organized as follows. Section II describes brief literature search related to this work. In section III Chromatic Balls algorithm is described with its drawbacks to show the problem part. Section IV describes the proposed algorithm with its both versions. The experimental setup and comparative results are provided in section V and VI. Finally the paper concludes in section VII. II. RELATED WORK A lot of research is headed in this direction for years by many authors. Detail analysis and literature search on this topic is done in my previous work [6]. Some of them introduced here. Bansal et al in 2004 [3] introduced the concept of