Abstract The paper proposed a method of hybridization of k-means algorithm and evolutionary programming. The blend of the two generates k number of clusters C = (c1, ..., ck) in the data space D = {x1, ..., xn}. These clusters will evolve in such a way that prediction of the upcoming trends of clusters in the application is possible. The proposed hybrid is named as evolutionary k-means clustering algorithm which is useful in generating and predicting clustering trends in an automated system. Keywords: Clustering, Data Mining, Evolutionary Programming, K-Means Clustering Trend Predictions using Evolutionary k-means Algorithm for Automated Clustering Jyoti Lakhani*, Dharmesh Harwani** * Department of Computer Science, Maharaja Ganga Singh University, Bikaner, Rajasthan, India. E-mail: jyotilakhanimgsu@gmail.com ** Department of Microbiology, Maharaja Ganga Singh University, Bikaner, Rajasthan, India. E-mail: dharmesh_harwani@hotmail.com 1. Introducion Clustering is a data mining technique of data analysis. Clustering is used to group objects on the basis of similarity; resulting similar type of objects in one cluster. The k-means algorithm is a clustering algorithm uses minimum distance to ind out the statistical clustering patterns. In this paper, the k-means algorithm is hybridized with the evolutionary programming to automate the process of inding clusters in a given data set. This is useful for detecting the up-coming probabilistic clusters for automated systems. 2. The k-means Algorithm The k-means algorithm is a clustering algorithm which deines clusters on the basis of minimum distance. From a Article can be accessed online at http://www.publishingindia.com given data points for k number of points which denote the centres of k clusters are selected randomly. By considering these points as centres of clusters, other points are also selected by following the minimum distance criteria (Gaussian distance). This process will be repeated until all clusters are created having all the data points with minimum distances. (MacQueen, 1967) 3. Evoluionary Programming Evolution is a continuous process of changing environment by selecting and ittest individuals from the population. Only the ittest will get the chance to reproduce and generate its successors. This process is called natural selection (Darwin, 1859). Based on the concept of natural selection and evolution, evolutionary programming is introduced by. Fogel in 1966 (Fogel, 1966). The classical evolutionary programming proposed by Fogel uses inite state automata for machine learning tasks such as prediction. To achieve the adaptive behaviour by a machine, prediction of the environment is must. The predictive behaviour is achieved by recombination and mutation operator. In a typical evolutionary programming, it is required to generate an initial population by random selection of the genetic content. Genetic content resides on the chromosomes and deines the behaviour of the individuals. A value of itness is assigned to the genetic content. In order to produce the next population the chromosome with highest itness value of genetic content is given a chance to reproduce offspring. The reproduction is implemented by one of the three possibilities copy, crossover and mutation. This process will follow that