Data Clustering using Genetic Algorithms

This project focus on the problem of data clustering, where the similarity between instances are measured by Euclidean distance metric. Here we include the case where the data dimension, the number of clusters and the size of data set could be very large. Genetic algorithm is used due to its capability to capture the global optimality, leading to a promising empirical performance in the given environment. Two crossing set selection methods, six crossover approaches and a fine-tuning technique are applied in the projects. Our experiments also verify that the genetic algorithm is computationally efficient, and achieve comparable performance , if not better, as the most commonly used state-of-the-art works on large data set.