Introducing
Your new presentation assistant.
Refine, enhance, and tailor your content, source relevant images, and edit visuals quicker than ever before.
Trending searches
Meta-learning
6
Meta-learning in predictive modeling
See our chapter on self organization of supervised models
Practical impacts
(in Modgen)
Meta-learning in recommender systems
Diversity in ensembles
Metadatabáze
Invitations
ceur-ws.org/Vol-1201/MetaSel2014-complete.pdf
BI Hackaton:
http://enterprise.hackathon.bi/
CIG meetings:
http://cig.fit.cvut.cz/doku.php?id=public:meetings:root
What is
meta-learning?
Our actual research
Predictive models
Definitions
Machine learning systems that utilize information from previous (related) problems (runs) to improve learning in future.
Jan Černý, CIG
Machine learning meta-data can be used
in several ways.
Clustering
Recommender systems
Ensemble methods
We will focus on ensemble methods
and knowledge base approaches.
Boosting
Several methods
Tomáš Bartoň, CIG
KNN and AR on LastFM data
Tomáš Řehořek, CIG
Automation in data preparation
http://predictorfactory.com/doku.php/algorithm
Bagging - bootstrap aggregating
Pavel Kordík
Jan Motl, CIG
Bias-variance decomposition of error
Knowledge based approaches
Meta-data
- Statistical:
- Statistical Measures (e.g. mean of numerical attributes)
- Simple Measures (e.g. number of attributes, classes )
- Information-based measures (e.g. entropy of classes)
- Histograms based
information regarding the distribution of values of attributes with relational nature (e.g. mutual information between symbolic attributes and class)
- Landmarking
use the performance of simple (fast) learners to predict the performance of candidate algorithms
http://expdb.cs.kuleuven.be/expdb/index.php
OpenML results for the BreastCancer UCI data
Company offering precise predictive models and recommender systems
StatLog project, Metal project: www.metal-kdd.org
Application field
- To select the suitable algorithm (algorithm recommendation)
- To combine algorithms in a clever way: ensembling
- To use the meta-information in order to construct better learning algorithm in general
Computational Intelligence research Group at CTU in Prague
Meta-learning in optimization
Faculty of Information Technology,
CTU in Prague
https://www.researchgate.net/publication/41942762_Meta-learning_approach_to_neural_network_optimization
Algorithm footprints: Kate Smith-Miles -Monash university, Australia
Meta-learning
Meta optimization for TSP problems
Oleg Kovářík, CIG