Research And Application Of Web Log Mining Based On Association Rules

This thesis aims to discuss the theory, algorithms and applications of web log mining Based on Association Rule.Firstly, this paper analyzes data mining and web mining roundly and puts emphases on the key technique in preprocessing of web log mining, especially introduces the data pretreatment process of web log mining.Secondly, We do a summary to definition,property,mining process,mining algorithms and present research of Association Rules Mining,and then analyzes the classical character of Apriori algorithm, finding out the disadvantage of the algorithm: this algorithm should scan the whole transaction database many times,which have much I/O overhead and can't achieve significant improvement of efficiency. and introduces one improved algorithms Imp_Apriori based on it, the main idea is: in the fact of associaton rules mining , the number of items is far less than the number of transactions, change the transaction database into item database,in which the item is regarded as index key ,it's record is the collection of transactions that includes the item , then the mining performs directly in the item database.Finally developed two test procedures based on Apriori algorithm and Imp_Apriori algorithm, through the same data sets in the same support and confidence level under the conditions of the two algorithms of their time digging in order to verify the feasibility of improving the algorithm.Finally,Based on the Web log mining theory and algorithm, design a system of web log mining,and put it in use of analyzing the Web log of "Enrollment Information Network",which is in the Network Center ,find out the uers'frequent access patterns,retrieve the data as for the evidence of the next works.