The Research On Apriori Algorithm Based On USI And Fundamentality Of Item

Data mining technology current is the research frontier with the informantion science field.It is an effective approach to resovle the problem of abundant data and scanty information. The paper studies mostly the problem of mining association rule.In many Data mining technology algorithms, mining association rule play one important role in the field.First of all, this thesis made an analysis and conclusion of the current situation and development of the association rule Data mining technology. Then is gave a comprehensive explanation of association rulei's basic knowledge. As many researchers have do a great number of studies and make remarkble achievments. So Abstract Apriori algorithm as a classic studie object. Firstly, clean up the general thinking of association rule algorithm. Conclude the current many main current improved thinking about association rule algorithm.Bring forward one new improved thinking about association rule algorithm:Base on the client being interested in candidate itemsets and the significant degree of itemsets thinking to improve Apriori algorithm.The main principle is abstract one candidate itemsets that consist of client being interested itemsets, then Scan the candidate itemsets, indicated the itemsets by matter-of-fact tag.Base the function of sustain degree to product frequent itemsets, so association rule can be draw-out.In the paper finality part, By contrasting between the new improved and the nomal Apriori algorithm, from the experiment data, conclude the efficiency of the new improved Apriori algorithm.