Step-by-Step Model for the Study of the Apriori Algorithm for Predictive Analysis

The goal of this paper was to develop an educational oriented application based on the Data Mining Apriori Algorithm which facilitates both the research and the study of data mining by graduate students. The application could be used to discover interesting patterns in the corpus of data and to measure the impact on the speed of execution as a function of problem constraints (value of support and confidence variables) or size of the transactional data-base. The paper presents a brief overview of the Apriori Algorithm, aspects about the implementation of the algorithm using a step-by-step process, a discussion of the education-oriented user interface and the process of data mining of a test transactional data base. The impact of some constraints on the speed of the algorithm is also experimentally measured without a systematic review of different approaches to increase execution speed. Possible applications of the implementation, as well as its limits, are briefly reviewed.