Anzeige
Erschienen in:
2019 | OriginalPaper | Buchkapitel
An Efficient MapReduce-Based Apriori-Like Algorithm for Mining Frequent Itemsets from Big Data
verfasst von : Ching-Ming Chao, Po-Zung Chen, Shih-Yang Yang, Cheng-Hung Yen
Erschienen in: Wireless Internet
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
Data mining can discover valuable information from large amounts of data so as to utilize this information to enhance personal or organizational competitiveness. Apriori is a classic algorithm for mining frequent itemsets. Recently, with rapid growth of the Internet as well as fast development of information and communications technology, the amount of data is augmented in an explosive fashion at a speed of tens of petabytes per day. These rapidly expensive data are characterized by huge amount, high speed, continuous arrival, real-time, and unpredictability. Traditional data mining algorithms are not applicable. Therefore, big data mining has become an important research issue.
Clouding computing is a key technique for big data. In this paper, we study the issue of applying cloud computing to mining frequent itemsets from big data. We propose a MapReduce-based Apriori-like frequent itemset mining algorithm called Apriori-MapReduce (abbreviated as AMR). The salient feature of AMR is that it deletes the items of itemsets lower than the minimum support from the transaction database. In such a way, it can greatly reduce the generation of candidate itemsets to avoid a memory shortage and an overload to I/O and CPU, so that a better mining efficiency can be achieved. Empirical studies show that the processing efficiency of the AMR algorithm is superior to that of another efficient MapReduce-based Apriori algorithm under various minimum supports and numbers of transactions.