An Efficient Association Rules Algorithm Based on Compressed Matrix

This paper analyses the classic Apriori algorithm as well as some disadvantages of the improved algorithms, based on which the paper improves the Boolean matrix. A row and a column are added on the former Boolean matrix to store the row vector of weight and account of the column vector. According to the quality of Apriori algorithm, Boolean matrix is largely compressed, which greatly reduces the complexity of space. At the same time, we adopt the method of weighting vector inner-product to find frequent Kitemsets so as to get the association rules. The complexity of space and time is developed to a large extent by the improved algorithm. In the end, the paper gives the computing procedure of the improved algorithm and by experiments, it proves that the algorithm is effective.