Analýza dat mechanických vlastností ocelových materiálů

This bachelor thesis deals with the application of knowledge discovery in databases on real data from the manufacturing industry area. The subject of the analysis is the output data from strength tests of steel materials. The work aims to find potentially usable recurring patterns, regularities, or irregularities in the analyzed data. The analysis is executed according to the basic phases of KDD that originate from the CRISP-DM methodology. Additionally, the analysis used the CF-Miner, KL-Miner and 4ft-Miner procedures implemented in the LISp-Miner system. The work is divided into the theoretical part and the practical part. The theoretical part presents the concept of KDD and its basic phases. Furthermore, the theoretical part presents the academic system for knowledge discovery in databases LISp-Miner and its procedures which were used for analysis. The division of the practical part corresponds with the basic phases of the KDD process. The practical part presents business understanding, data understanding, and data preparation. It is followed by a description of the modeling and selected hypotheses. The analysis concludes by summarizing the results of each analytical task. Although the data used was slightly unsuitable for analysis, the thesis succeeded in finding potentially useable recurring patterns, regularities and irregularities in the analyzed data.