What's new



- Jan/2005  attending The first Japanese German Frontiers of Science Symposium as a general participant (without talk, but poster presentation)


- Nov/2004   BEST IMPLEMENTATION AWARD of    FIMI04(Frequent Itemset Mining Implementations)

   "LCM v.2: Efficient Mining Algorithms for Frequent/Closed/Maximal Itemsets"

   Takeaki Uno, Masashi Kiyomi and Hiroki Arimura

FIMI is a workshop of competition on the implementations of frequnet itemset mining problems. Frequent itemset mining is, roughly speaking, to find combinations of items which are frequently purchased together, from database of shops. This problem is very fundermental in data mining, and has been studied actively. It has many applications, so there are many implementations. On this competition, the speed of existing and new implementations and compared, by computational experiments. Our implementations won at over half instances, and got the best implementation award.

FIMI_award_s.jpg (5352 バイト)


FIMI_prize_s.jpg (6628 バイト)

Prize. It is the "most frequent itemset"



- Mar/2004  DEWS(15th Data Engineering Workshop) Best Paper Award

   "Fast Mining Scheme for Un-ordered Tree Patterns from Semi-Structured database"

   Shinji Fusanobu, Tatsuya Asai, Hiroki Arimura, Takeaki Uno, Shin-ichi Nakano

A research of fast algorithms for mining all frequent appearing (un-ordered) tree patterns from semi-structured database. We developed efficient algorithms by combinating an efficient enumeration algorithm of un-ordered trees and an occurrence computing algorithm.


- Oct/2003

A student of doctoral course came to my laboratory, who is belonging to The Graduate University for Advanced Studies, Department of InformaticsAdvanced.


 Uno's Homepage          NII Homepage