18.11.2012 Views

anytime algorithms for learning anytime classifiers saher ... - Technion

anytime algorithms for learning anytime classifiers saher ... - Technion

anytime algorithms for learning anytime classifiers saher ... - Technion

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Technion</strong> - Computer Science Department - Ph.D. Thesis PHD-2008-12 - 2008<br />

7. Corral: An artificial dataset first used by John, Kohavi, and Pfleger (1994).<br />

8. Glass: In this domain, taken from the UCI Repository, the goal is to determine<br />

the type of glass from its characteristics.<br />

9. Heart: In this domain, taken from the UCI Repository, the goal field refers<br />

to the presence of heart disease in the patient. It is integer valued from 0<br />

(no presence) to 4.<br />

10. Hepatitis: In this problem, taken from the UCI Repository, the diagnosis is<br />

known, and the problem is to determine the likely outcome of the disease.<br />

11. Iris: This dataset is taken from the UCI Repository and contains 3 classes<br />

of 50 instances each, where each class refers to a type of iris plant.<br />

12. Monks problems: This set, taken from the UCI Repository, contains three<br />

problems. Each example is represented by 5 nominal attributes in the range<br />

1, 2, 3, 4. The problems are:<br />

• Monks-1: (a1 = a2)or(a5 = 1).<br />

• Monks-2: exactly two of (a1 = 1, a2 = 1, a3 = 1, a4 = 1, a5 = 1).<br />

• Monks-3: ((a5 = 3)and(a4 = 1))or((a5 �= 4)and(a2 �= 3)), with an<br />

added 5% class noise.<br />

The original datasets are already partitioned into training and testing sets.<br />

13. Mushroom: This dataset, taken from the UCI Repository, includes descriptions<br />

of hypothetical samples corresponding to 23 species of gilled mushrooms<br />

in the Agaricus and Lepiota family. Each species is identified as<br />

edible or poisonous.<br />

14. Nursery: This database, taken from the UCI Repository, was derived from<br />

a hierarchical decision model originally developed to rank applications <strong>for</strong><br />

nursery schools.<br />

15. Pima: The Pima Indians Diabetes dataset, taken from the UCI Repository,<br />

includes several medical tests and the possible classes are either the patient<br />

is healthy or the patient has diabetes.<br />

16. Solar Flare: This problem is taken from the UCI Repository. Each instance<br />

represents captured features <strong>for</strong> one active region on the sun. Among the<br />

three possible classes, we considered the C-class flare where the instances<br />

are more distributed.<br />

140

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!