Veri madenciliği ve Anadolu Üniversitesi açıköğretim sisteminde bir uygulama

In Data mining, the discovery of frequent sets of items that occur together in a dataset is a fundamental task, particularly in transaction datasets. This thesis focus on closed frequent itemsets, a compressed representation of frequent patterns. Since mining all frequent itemsets generates duplicates and subsets, and with dense dataset, it gives a huge number of generated frequent itemsets which may be not understandable, then to avoid generate duplicates and subsets, and to get useful and compressed information about data under study, closed frequent itemset technique was adopted in this study. In this thesis, the CHARM algorithm to discover closed frequent itemsets is proposed to identify frequent itemsets for Anadolu University Open Education System data to find out the collection of books taken together. The reason of choosing CHARM algorithm in this study is that it employs a vertical layout which facilitates the process of computing support count, moreover, it employs a novel way to traverse the search space. The solution is implemented in R software environment for statistical computing and graphics. The result obtained shows that the encoded CHARM algorithm can get an efficient performance. Moreover, the result obtained could be useful to help decision makers to package the books in packets to achieve the aim of the study.