The traditional algorithms for mining association frequent patterns generate conditional sub tables

which costs much runtime and memory space.To solve these problems

a new algorithm FIA(Frequent Iternset Algorithm)is proposed. The FIA algorithm adopts a binary of symbols to compress the store data.The algorithm using logic of symbols to express data in a database

which only after one scan

and establish a binary vector and upper triangular frequent matrix

according to the both to produce a set of frequently.Thereby effectively narrowing the search space

speed up the processing speed.Through analysis showed that

FIA algorithm more effective than Apriori algorithm.