An Improved Apriori Algorithm Based on Matrix Compression
Tong Liu, Xiaopeng Ji, Yongquan Yang, Wei Zhi-qiang · International Symposium on Computational Intelligence and Design · 2013
As a classical algorithm using association rules in data mining, the Apriori algorithm has the defect of producing a large number of candidate item sets and scanning the database many times. This paper puts forward an improved MC-Apriori algorithm based on matrix compression, which scans a database and turns it into a Boolean transaction matrix, and then compresses the transaction matrix according to the relevant properties to reduce the amount of computation. The experimental results show that the MC-Apriori algorithm performance has been significantly improved.