Analysis of library loan records by association rules: Düzce university case study

Information Technologies has become widely used in many institutions and companies due to the cheapening of its costs. In this way, the amount of data generated and stored in various systems has reached very large volumes. Data Mining, which can described as the processes of extracting meaningful information from these raw data, has gained importance in recent years. In this study, the discovery of association rules was carried out by running Apriori, Fp-Growth and Eclat algorithms, which are one of the Data Mining methods, on the Düzce University Library Records dataset, according to various support and confidence values. No examples of repetitive rules were found at the book level. However, at the level of the book categories in the library, patterns of borrowing together were discovered and the rules related to them were revealed. At the same time, the effect of the gender of the readers on their book loan behavior was also investigated and the results were evaluated. In addition to finding these rules, the performances of the specified Association Rules algorithms on Python, R and SPMF platforms were compared on the basis of their running time (ms) and the amount of memory they occupied (MB) during the experiments. According to the results obtained, it was seen that the SPMF program stood out in the men's data set with the least memory occupation and the lowest running time values.