摘要
The content-ignorant clustering method takes advantages in time complexity and space complexity than the content based methods.In this paper,the authors introduce a unified expanding method for content-ignorant web page clustering by mining the "click-through" log,which tries to solve the problem that the "click-through" log is sparse.The relationship between two nodes which have been expanded is also defined and optimized.Analysis and experiment show that the performance of the new method has improved,by the comparison with the standard content-ignorant method.The new method can also work without iterative clustering.
The content-ignorant clustering method takes advantages in time complexity and space complexity than the content based methods. In this paper, the authors introduce a unified expanding method for content-ignorant web page clustering by mining the ＂click-through＂ log, which tries to solve the problem that the ＂click-through＂ log is sparse. The relationship between two nodes which have been expanded is also defined and optimized. Analysis and experiment show that the performance of the new method has improved, by the comparison with the standard content-ignorant method. The new method can also work without iterative clustering.
