| Association rules have been regarded as a very important topic of data mining research. Since Agrawal proposed the concept of association rules and the first association rules mining algorithm, that is Apriori, a lot of researchers have broadly researched on Apriori algorithm because it has commercial values and theoretic values. On this basis, many new association rules mining algorithms are proposed by optimizing and improving Apriori algorithm continuously in order to impove the efficiency of data mining.Whereas the efficiencies of these algorithms are enhanced, there remain some deficiencies in these algorithms. In addition, there are two prevalent problems in association rules mining: How to acquire the desired results efficiently and immediately when the mining data updates constantly? Usually, it's necessary to set some parameters for customers before mining, and mostly they have to adjust these parameters many times to acquire the satisfactory rules, thus how to calculate efficiently during the repetitious process? The main purpose for proposing the incremental updating algorithms is to solve these problems.To solve the first problem, based on some researches about the incremental updating algorithms, with regard to the deficiencies that FUP algorithm need scan database many times and NEWFUP algorithm does't consider the cost of producing the mothball frequent item sets, a new incremental updating algorithm using mothball frequent item sets which is called UMMFUP is proposed to be used in the case of adding new data to the database. The feasibility and efficiency of UMMFUP algorithm is demonstrated using an experiment. At the same time, UMMFUP2 algorithm using mothball frequent Item sets is proposed to be employed in the case of deleting the data from the database. Also an experiment is made to validate UMMFUP2 algorithm. The result reflects that the times of reusing mothball frequent item sets are more and the feasibility and efficiency of UMMFUP2 Algorithm is better. |
PDF Full Text Request