23.06 14.30 IDSIA - Nicolaj Tatti
by Announcements of talks＠IDSIA
It will start in about 10 minutes, room 222 Galleria 1.
**
Monday 23 July 2012, 14.30, Galleria 1, Manno
Room 222, DSAN
SPEAKER: Nikolaj Tatti, University of Antwerp
http://adrem.ua.ac.be/ntatti
TITLE: Discovering Descriptive Tile Trees
When analysing binary data, the ease at which one can interpret results
is very important. Many existing methods, however, discover either
models that are difficult to read, or return so many results
interpretation becomes impossible. Here, we study a fully automated
approach for mining easily interpretable models for binary data. We
model data hierarchically with noisy tiles---rectangles with
significantly different density than their parent tile. To identify good
trees, we employ the Minimum Description Length principle.
We propose Stijl, a greedy any-time algorithm for mining good tile trees
from binary data. Iteratively, it finds the locally optimal addition to
the current tree, allowing overlap with tiles of the same parent. A
major result of this paper is that we find the optimal tile in only
Theta(NM min(N,M)) time. Stijl can either be employed as a top-k miner,
or by MDL we can identify the tree that describes the data best.
Experiments show we find succinct models that accurately summarise the
data, and, by their hierarchical property are easily interpretable.
BIO
Nikolay Tatti has got his Phd at the Helsinki University of Technology.
The title of his thesis was "Advances in Data Mining: Itemsets as
Summaries". Itemset analysis and pattern mining is indeed his main topic
of research. Currently, he is working at the University of Antwerp.
During the last years he has had constant presence in the top
data mining conferences, such as KDD, ICDM, SDM and ECML PKDD.
12 years, 5 months
23.06 14.30 IDSIA - Nicolaj Tatti
by Announcements of talks＠IDSIA
I have added the room and the speaker bio.
best, Giorgio
**
Monday 23 July 2012, 14.30, Galleria 1, Manno
Room 222, DSAN
SPEAKER: Nikolaj Tatti, University of Antwerp
http://adrem.ua.ac.be/ntatti
TITLE: Discovering Descriptive Tile Trees
When analysing binary data, the ease at which one can interpret results
is very important. Many existing methods, however, discover either
models that are difficult to read, or return so many results
interpretation becomes impossible. Here, we study a fully automated
approach for mining easily interpretable models for binary data. We
model data hierarchically with noisy tiles---rectangles with
significantly different density than their parent tile. To identify good
trees, we employ the Minimum Description Length principle.
We propose Stijl, a greedy any-time algorithm for mining good tile trees
from binary data. Iteratively, it finds the locally optimal addition to
the current tree, allowing overlap with tiles of the same parent. A
major result of this paper is that we find the optimal tile in only
Theta(NM min(N,M)) time. Stijl can either be employed as a top-k miner,
or by MDL we can identify the tree that describes the data best.
Experiments show we find succinct models that accurately summarise the
data, and, by their hierarchical property are easily interpretable.
BIO
Nikolay Tatti has got his Phd at the Helsinki University of Technology.
The title of his thesis was "Advances in Data Mining: Itemsets as
Summaries". Itemset analysis and pattern mining is indeed his main topic
of research. Currently, he is working at the University of Antwerp.
During the last years he has had constant presence in the top
data mining conferences, such as KDD, ICDM, SDM and ECML PKDD.
12 years, 5 months
23.06 14.30 IDSIA - Nicolaj Tatti
by Announcements of talks＠IDSIA
Monday 23 July 2012, 14.30, Galleria 1, Manno
Room to be announced
SPEAKER: Nikolaj Tatti, University of Antwerp
http://adrem.ua.ac.be/ntatti
TITLE: Discovering Descriptive Tile Trees
When analysing binary data, the ease at which one can interpret results
is very important. Many existing methods, however, discover either
models that are difficult to read, or return so many results
interpretation becomes impossible. Here, we study a fully automated
approach for mining easily interpretable models for binary data. We
model data hierarchically with noisy tiles---rectangles with
significantly different density than their parent tile. To identify good
trees, we employ the Minimum Description Length principle.
We propose Stijl, a greedy any-time algorithm for mining good tile trees
from binary data. Iteratively, it finds the locally optimal addition to
the current tree, allowing overlap with tiles of the same parent. A
major result of this paper is that we find the optimal tile in only
Theta(NM min(N,M)) time. Stijl can either be employed as a top-k miner,
or by MDL we can identify the tree that describes the data best.
Experiments show we find succinct models that accurately summarise the
data, and, by their hierarchical property are easily interpretable.
12 years, 5 months