The research of association rules mining algorithm based on binary

Abstract

An algorithm of association rules mining based on binary has been introduced to solve two problems that how to easily generate candidate frequent itemsets and fast compute support. However the basic notion of presented algorithms in generating candidate itemsets is still similar to Apriori. In some degree the efficiency of these algorithms is very confined, and so this paper proposes two different searching strategies of association rules mining algorithms based on binary, which are suitable for mining corresponding characteristic database. One is that the notion of generating candidate frequent itemsets is similar to up-down searching of traditional association rules mining algorithm, which uses the method of forming subset to generate candidate frequent itemsets from long to short and is suitable for mining long frequent itemsets. The other is that the method of increasing value is used to generate candidate frequent itemsets, which is more efficient than Apriori based on binary and is more suitable for mining short frequent itemsets, in this mining course length of candidate frequent itemsets crossways varies from short to long. The both algorithms use digital character to reduce the number of scanned transactions. The experiment based on above three algorithms indicates that the efficiency of two presented algorithms is fast and efficient when mining corresponding characteristic database.

The application of a top-down algorithm in neighboring class set mining

“…Namely, it uses digit logical operation to generate k-candidate frequent itemsets of digit form by (k+1)-digit non-frequent itemsets. The algorithm adopts search strategy which is similar to B_UDMA as in [7] and ITDASN as in [8]. The course of generating is expressed as follows:…”

A Vector Operation Based Fast Association Rules Mining Algorithm

“…Table IV shows the variation of the execution time with their correspondence support. Since these two algorithms execute on the same test data, it is can be clearly seen that the efficiency of the V_Apriori algorithm is better than the classical Apriori algorithm [9]. (2) Space(Capacity)complex analysis The Apriori algorithm produces a mass of candidate itemsets while executing, and all these sets need to be stored in the main memory of the PC so that they can be used in the prune step to generate the frequent itemsets.…”

An Improved Top-Down Data Mining Algorithm for Long Frequents

“…COMPARING CAPABILITY OF THESE ALGORITHMS Aiming to B_Apriori represented as in [3] and B_UDMA represented as in [5], this paper proposes an improved top-down association rules mining algorithm based on sequence number (ITDASN). We compare the method of computing support used by these algorithms which are expressed as follows: B_Apriori: The algorithm generates candidate frequent itemsets which contain the number of items from fewness to many, namely, the algorithm uses up search strategy to generate candidate, which is suitable for mining short frequent itemsets.…”

“…In order to solve the first problem, some algorithms were presented, such as DMFI [1] and DMFIA [2] . And then, in order to solve the second problem, some algorithms based on binary were presented, such as B_Apriori [3] , B_ARDSM [4] and B_UDMA [5] . The kind of algorithm computes support by binary logic operation to improve efficiency of algorithm.…”