Name (abbrev)

Name (full)

Category

Last update

UTUBE

Modelling, Diagnosis, Control

b D, Y

Application domain

Further specifications

Learning qualitative models from example behaviors

dataset = data + background theory

Type

Format

Complexity

ILP

Prolog (mFOIL format)

4 positive and 543 negative examples

WWW / FTP

ftp://ftp.mlnet.org/ml-archive/ILP/public/data/utube/

Contact person(s)

Related group(s)

Optional contact address

2. Dzeroski, Saso

2. Jozef Stefan Institute, Department of Intelligent

References

Lavrac, N., Dzeroski,S.
Inductive Logic Programming: Techniques and Applications.
Ellis Horwood, Chichester, 1994. Chapter 13.

Annotations

The dataset is intended to learn a QSIM type qualitative model of two connected containerns - a dynamic system, known also as the U-tube. The target predicate legalstate/4 has as arguments the qualitative values of the four system variables (water levels and their rates of change for each of the two containers). Positive facts corresponding to this predicate describe the states that appear in a qualitative behavior of the U-tube (4 facts). Negative facts (543) characterize those states that are impossible in such a system. Background knowledge defines the qualitative constraints of the QSIM formalism: add/3, multiply/3, deriv/2, m_plus/2, m_minus/2, and minus/2. The achieved results are described in [Lavrac 94].

Comments