Dealing with temporality when inducing association rules from a retail database

One of the main motivations for employing database mining relates to the decision support problems faced by many retail organizations, particularly when the temporal aspect of the data should be taken into account to support management decisions. The focus of this work was the proposal and development of a computational system able to identify patterns in a database containing sale records from a retail organization and, based on them, infer association rules that also deal with the temporal aspect of the data. Experiments with the system on real data from a retail company, after a two-step pre-processing of the raw data, are presented. The obtained results are analysed and a few conclusions are drawn.