Market basket analysis in data mining and finding association rules

Today, large amounts of data can be collected and stored by using technology. However, there is a necessity of automatic analysis using computer technology and computer programmes which is developing day by day in order to analyze the data, that is difficult to be analyzed by manuel and can not be seen. Making summaries in the simple way by finding patterns, tendencies, anormalities from the database is one the most common thing in the information age. Data mining is the process of finding the rules and the correlations among the large amounts of data by the computer programmes, which are understandable, potentially useful and provide predictions about the future. The utilization of data mining in a wide selection of fields is increasing. One of the areas is the market-basket analysis that is to have the rules and associations from the data about customer, products and sales. In this analysis, gathering the association rules-one of the subjects in the data mining- and having the sales relationships between the products are two factors of increasing rate of profit in the companies. Association rules provide predictions about the future by discovering relations between the objects which act together in the transactional sales data and the objects. Lots of algorithms has been developed since the beginnings of 1990?s. These algorithms have different working methods and different superiorities on each other in the different conditions. The common logic of these algorithms is that passing over the database, combining, pruning and finding the association rules between the items by using the minimum support threshold.In this thesis, concepts about the data mining and basic algorithms especially using in the market-basket analysis to produce the association rules are examined in details and compared with each other. Also, an application is developed to find association rules from sample datasets by using two different algorithms.Keywords: Data mining, Market-basket analysis, Association rules, Association rule mining algorithms, Apriori algorithm, FP-Growth algorithm.