Clustering categorical datasets using multi objective genetic algorithm

The process of classifying data in computer science has been the interest of lots of practices in recent years and lots of techniques have been suggested about this topic. This topic concentrates on how to classify these data appropriately using their attribute values. As a result of this, data which are similar can be put in same or near clusters and there can be made some inferences from these clustering results. These area can deepen with the status of data sets. For example classification of large and small scale data specialize very differently.Multi objective clustering as a kind of clustering is a trending topic in recent years. This type of clustering takes into account more than one objective and gives clustering results of data as an output.In this thesis, it has been studied on basically how to classify categorical data sets with a multi objective structure. Obtained results have been compared with the results of k-mod algorithm which do the clustering process with one objective. Using the purity measure, it has been shown that in every condition, the proposed procedure gives better clustering results.In further steps of the work, the problems occurring with the clustering of large scale data investigated and some solutions have been proposed for them. For these conditions, the procedure that has been proposed is to generate frequent item sets from the data and classifying these frequent item sets efficiently. In the end, outputs that have been gathered illustrate the efficiency and consistency of this procedure.