| Summary: | Big Data era is currently generating tremendous
amount of data in various fields such as finance,
social media, transportation and medicine. Handling and processing this “big data” demand powerful data mining methods and analysis tools that can turn data into useful knowledge. One of data mining methods is frequent itemset mining that has been implemented in real world applications, such as identifying buying
patterns in grocery and online customers’ behavior.Apriori is a classical algorithm in frequent itemset mining, that able to discover large number or itemset with a certain threshold value. However, the algorithm suffers from scanning time problem while generating candidates of frequent itemsets.This study presents a comparative study between several Apriori-variant algorithms and examines their scanning time.We performed experiments using several sets of different transactional data.The result shows that the improved Apriori algorithm manage to
produce itemsets faster than the original Apriori
algorithm.
|
|---|