The Research Based On Time Series And Maximum Clique For Data Mining Of Association Rules

The association rule mining is one of the most active research techniques in the data mining. It was most early proposed by Agrawal (in 1993). At first, the proposed motive aims at the question of shopping basket analysis (Basket Analysis), and its goal is to discover the different commodity relation rule between the transaction database (Transaction Database). The transaction database may save customer's related transactions (for example shopping project and so on). By these data intelligent analyses, we may obtain the general rule concerned customer to purchase the pattern. These rules have portrayed customer's purchase behavior pattern, we may use them to instruct the merchant scientifically to arrange to import goods, the stock as well as goods design and so on. The association rule also may obtain the widespread discussion in other domains, like the table of contents design, the commodity advertisement mails the analysis, supplements sales, the warehouse plan, the network fault analysis, the market rule, the advertisement plans, the classified design and so on. The association knowledge (Association) reflects the dependence or the association between an event and the other event, the association may divide into the simple association, the time series (Time Series) association, the causes and effects association, the quantity association and so on. These associations are not always knew beforehand, but they are the data association analysis obtains through the database, thus they have the new value to the business decision.In fact, we suppose the association rules obtained by the majority algorithm are effective forever, but the time is the important attribute in the real world, time attribute of the large capacity data sets is possibly very essential for the users. What the users care often is the some region data but not the entire data for a while, but the data in the specific time region possibly causes the association rule during the specific data. The method of solving this question is the consideration time factor in the algorithm, therefore the table must include three fields at least —transaction number, the tense sector and the items in the database. The tense sector here has reflected the corresponding item sequence which occurs or the time scope which is collected. The association rule mining may be pretreatment using the tense restrain and so on and filter the time interval data that is not concerned by the users. Filtering the database is to reduce the scanning space, to reduce the input and output price, the reduced memory demand, then enhances the key...