A Kind of Improved Data Clustering Algorithm in Web Log Mining
Jin Guo, Shengbing Zhang, Zheng Qiu · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2015
Aiming at the user clustering and page clustering in Web log mining and based on the analysis of K-means clustering algorithm and matrix clustering algorithm, the paper presented an improved clustering algorithm that combining fuzzy matrix algorithm with K-means algorithm.Extract compressed sub-matrix from relational matrix of user and page, establishing user interval, and then divide all users into large intervals and separate the noise data, obtain initial value and classified number for K-means algorithm, effectively solve the defect in the K-means algorithm that always suppose or make a try to definite the classified number and the initial value, also include the lacking to exclude the noise data obstruction.