A Comparative Study of Pattern Recognition Algorithms on Sales Data

Abstract

In the realm of Data Mining looking for patterns and association rules is a very critical task and has been widely studied in the past years. There exist several data mining algorithms to find Association Rules in given datasets. One of the most popular and widely used algorithm is the Apriori algorithm to find patterns and itemsets in huge datasets and getting the association rules between them. This is done to gather knowledge from otherwise unsuspecting and random data. The Fp-Growth algorithm is similarly a different algorithm which uses an extended frequent pattern prefix-tree data structure for storing critical data after compression about frequent pairs. In this paper we do a comparative analysis of the 2 most popular pattern recognition algorithms and their performance on sales data of a college canteen sales transnational database where each record consists of items purchased by customer.

Data mining for road accident analysis in a big data context

“…Many algorithms [18] have been designed for generating association rules, we can cite that APRIORI and FP-Growth, are the two most popular for finding association rules in data sets, Moreover, they prove their performance on a transactional database where each transaction design a pattern (an itemset). In [19] a study shows that FP-growth provides much more consistent and quicker performance then APRIORI, because that FP-growth uses divide and conquer strategy and it needs only two passes over the datasets so it is much more scalable.…”