International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 04 Issue: 04 | Apr -2017 www.irjet.net p-ISSN: 2395-0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 2363 K-Means algorithm with different distance metrics in spatial data mining with uses of NetBeans IDE 8.2 Ms. Kothariya Arzoo 1 , Asst. Prof. Kirit Rathod 2 1 M. Tech student, Computer Engineering, C.U. Shah College of engineering and technology, Gujarat, India 2 Asst. Prof , Computer Engineering, C.U. Shah College of engineering and technology, Gujarat, India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract -Data mining is a process of finding useful information from large database. Clustering is a process of grouping the same characteristics elements in one group (cluster) and while distinct characteristics elements in different group (cluster).K-Means is very simple and very popular clustering technique. In this paper we will do the experiments with spatial data mining. Spatial data mining is the application of data mining. In spatial data mining spatial or geographic dataset is used. Distance metrics play very important role in clustering technique. In this paper we will do the experiments with the NetBeans IDE 8.2 and taking spatial data from Indian government website. This paper includes implementation analysis on k means clustering with different distance metrics with taking spatial database. Key Words: Clustering, Spatial Data mining, K-Means, Distance Metrics 1. INTRODUCTION Data mining [1] [2] involves process of finding useful information. Data mining consists of steps of a) data cleaning, b) data selection, c) data integration, d) data mining, e) pattern evaluation and finally f) knowledge representation [2]. Nowadays large amount of data generating day by day but this all data are not useful for us. So. Data mining is very essential for us in daily life. The data Clustering [3] involves the process of dividing the same data in one cluster and distinct data in different cluster so that inter cluster property should high and intra cluster property should low. These two are the clustering property. Clustering is the popular research topic nowadays. Classification of Clustering algorithm as: partitioning based clustering, hierarchical based clustering, grid based clustering etc [4]. A partitional clustering a simply a division of the set of data objects into non-overlapping subsets such that each data object is in exactly one subset. • A hierarchical clustering is a set of nested clusters that are structured as a tree. Clustering process involves feature selection, clustering algorithm, cluster validation, result interpretation, knowledge [5]. K- means [6] [7] is very simple clustering algorithm in data mining. Clustering is unsupervised technique [3] [4] such that there is no any test data and training data are available by which we can predict our result. Clustering is totally unsupervised method for data mining. There are many research and review paper available for k means and its variants [6]. There are many parameters for modifying k means algorithm based centroid initialization, distance metrics, improving accuracy. There are many limitations of k means algorithm [8] like handling empty cluster, outlier detection, distance metrics, number of cluster predefined, cluster center chosen randomly etc. Spatial data mining [9] [10] [11] requires specific resources to get the spatial database in specific format. The application covered by spatial data mining are geomarketing, environmental studies, risk analysis, and so on. The aim of clustering is to automatically find groups of instances that are similar to each other. For example, in classroom there is cluster of students with similarities in their birth date is in same month. Using GIS, the user can query spatial data and perform simple analytical tasks using programs or queries. However, GIS are not designed to perform complex data analysis or knowledge discovery [12]. There are many distance metrics [13] [14] [15] are available like Euclidean, Manhattan (city block), Chebychev, Minkowski etc. 1.1 K-Means Algorithm The K-means [6] [7] algorithm involves randomly selecting K initial centroids or mean where K is a user defined number of desired clusters. For each of the object the distance is calculated between center points and data points and with minimum distance data, cluster is generated. This data points are far from another cluster or group. This computation is stop when no center points do not move any more. K means algorithm work as follows: 1. Initialization by setting initial centroids with a predefined k. 2. Cluster or group the data points in given k clusters. 3. Assign data or objects to nearest cluster center as per distance function. 4. When all objects are assigned recalculate or update the position of k centroids. 5. Repeat step 3 and 4 until the centroids no longer move. Diagrammatic representation of k means algorithm as follows [17]: