1. IJCNN 2012 Competition: Classification of Psychiatric Problems Based on Saccades Włodzisław Duch 1,2, Tomasz Piotrowski 1 and Edward Gorzelańczyk 3 1 Department of Informatics, Nicolaus Copernicus University, Toruń, Poland 2 School of Computer Engineering, Nanyang Technological University, Singapore 3 Department of Informatics, Nicolaus Copernicus University, Toruń, Poland

2. Are we doing the right thing? Perhaps simplest approach is sufficient?

3. Perhaps the data is trivial?

4. Data mining packages No free lunch => provide different type of tools for knowledge discovery: decision tree, neural, neurofuzzy, similarity-based, SVM, committees, tools for visualization of data. Support the process of knowledge discovery/model building and evaluating, organizing it into projects. Many other interesting DM packages of this sort exists: Weka, Yale, Orange, Knime... >170 packages on the-data-mine.com list! We are building Intemi, radically new DM tools. GhostMiner, data mining tools from our lab + Fujitsu: http://www.fqspl.com.pl/ghostminer/ http://www.fqspl.com.pl/ghostminer/ Separate the process of model building (hackers) and knowledge discovery, from model use (lamers) => GM Developer & Analyzer

5. What DM packages do? Hundreds of components... transforming, visualizing... Rapid Miner 5.2, type and # components: total 712 (March 2012) Process control 38 Data transformations 114 Data modeling 263 Performance evaluation 31 Other packages 266 Text, series, web... specific transformations, visualization, presentation, plugin extensions... ~ billions of models! Visual “knowledge flow” to link components, or script languages (XML) to define complex experiments.

7. Test the simplest methods with O(nd) complexity & use them first in meta-learning.Test the simplest methods with O(nd) complexity & use them first in meta-learning. Provide reference results.Provide reference results. Compare with state-of-the-art results.Compare with state-of-the-art results. Identify trivial cases.Identify trivial cases. Provide reference results.Provide reference results. Compare with state-of-the-art results.Compare with state-of-the-art results. Identify trivial cases.Identify trivial cases.

8. Six O(nd) classifiers (n = # samples, d= dimensionality). 1.MC, Majority Classifier.

9. 2. 1NP, Nearest Prototype classifier Prototypes taken from class means, Euclidean distance only.

10. Thank you for lending your ears! Thank you for lending your ears! Google: W. Duch => Papers & presentations & more