Recent Papers
The notion of contrast set mining has been applied to a correlation mining algorithm so that it can detect even non-characteristic and therefore implicitly valuable pairs of itemsets showing increases in their correlations.
Discovery of Hidden Correlations in a Local
Transaction Database based on Differences of Correlations
Taniguchi, T and Haraguchi, M
Engineering Application of Artificial Intelligence, 19(4), pp. 419 - 428, Elsevier, 2006.
Are you looking for interesting documents or pages?
To do so, the paper below presents an efficient algorithm for finding the best N numbers of clusters of similar pages including even potentially interesting pages with lower ranks.
An Extended Branch and Bound Search Algorithm
for Finding Top-N Formal Concepts of Documents
Haraguchi, M. and Okubo, Y.
LLLL06 21 April 2006
to appear as a selected paper in "New Frontiers in Artificial Intelligence", Springer LNAI
A related paper, written by Okubo and Haraguchi, will be presented at IEEE/WIC/ACM WI-2006 Web Intelligence, held in Hong Kong, this Dec.
The above LLLL06 paper has been issued from the motivation stated in this paper:
A Method for Pinpoint Clustering of Web Pages with Pseudo-Clique Search
(Note: this href reference lead you to the Springer-Link page)
Haraguchi,M. and Okubo,Y.
Lecture Notes in Computer Science, Volume 3847, pp. 59-78 (2006)