Data mining na reálných datech

The aim of this bachelor's thesis is to find interesting and potentially useful relations on real data or on the other way confirming known relations using data mining and the LISp-Miner system. The analysis was based on the dataset with files of video game digital distributor service called Steam developed by Valve company. Dataset was downloaded from the server kaggle.com which focuses on data processing. The whole work is divided into theoretical and practical part.The theoretical part describes what is knowledge discovery in databases, evolve of this field and its methodologies. Then, there are described particular phases of the whole KDD process, according to CRISP-DM methodology. After that, thesis introduces three pillars of KDD: databases, statistics and machine learning. In the end of theoretical part is presented LISp-Miner system which will be used during the practical part. It continues with GUHA method and particular used procedures CF-Miner, MCluster-Miner and KL-Miner.The practical part follows the process of the KDD step by step, exactly by the description in the theoretical part. It begins with description about data acquisition, continues with data pre-processing, formulating of the analytical questions and performing the analysis. The thesis ends with evaluating and interpreting results of the analysis.