Multidimensional Association Rules

Data mining is one of the most active fields in database. Because of its wide application background and realistic significance, the technique and application of data mining advance rapidly and draw upon the attention of the international academic circle. Mining association rules is an important research aspect in the field of the data mining . There are many types of association rules, while mining multi-dimensional association rules is a new field at present . It takes not only association among items into account, but also the event surrounding information (eg., time and place). Consequently, it injects new meanings into traditional rules mined, supplies more exact reference information, and also bring forward new challenge. The text works over multi-dimensional association rules.It mainly includes two aspects. The first part is the basic arithmetic MB of mining . It adopts the extending item method and the dimensions hold the line . Except the last element of items, the other keeps the same, thus high rank candidate items are created. In order to improve efficiency, we use a data structure ---index tree. After creating high rank candidate, we set up the index tree for it .Dimensions are branches and leaves are made of items On scaning database when reading an affair, look up the index trees. If matching , each corresponding item count plus one. The second part is Vector mining arithmetic:VM. The process of the arithmetic is described as follows. We use hash technique to reduce generating candidate twe-items. On the back phase, the sets of candidate is very small. We use vector to represent frequent twe-items. On joining those two-items, it can generate all candidate items. After scaning database once, we can get all frequent items. Then the Vector mining arithmetic saves 110 time.