Abstract：A new algorithm based on double compression, which was called as Apriori double compres- sion （Apriori_ DC ） , was proposed, according to the performance bottleneck problem of Apriori algorithm. Two ways were used to improve performance： the transaction database was continually compressed to re- duce the transaction record and the total item in the database ; the number of the next candidate frequent item set was to reduce by compressing the frequent item set. The experiments showed that Apriori_ DC al- gorithm had better performance than Apriori algorithm when the support ratio was the same and the record number of the database was different or the record number of the database was the same and the support ratio was different. The experiment also showed that the record number of the database was continually reduced during the execution of the Apriori_ DC algorithm.
2017, Vol. 30 
