The minimal-model semantics of causation is an essential concept for the identification of a best ffitting model in the sense of satisfactory consistent with the given data and be the simpler, less expressive model. Therefore to develop an algorithm being able to derive a minimal model is an interesting topic in the area of causal model discovery. various causal induction algorithms and tools developed so far can not guarantee that the derived model is a minimal model. This paper proves that the MML induction approach introduced by Wallace, et al is a minimal causal model learner. The experimental results obtained from the tests on a number of both artificial and real models provided in this paper conform this theoretical result.
[1]
Judea Pearl,et al.
ON THE CONNECTION BETWEEN THE COMPLEXITY AND CREDIBILITY OF INFERRED MODELS
,
1978
.
[2]
Peter M. Blau,et al.
The American Occupational Structure
,
1967
.
[3]
Kevin B. Korb,et al.
Causal Discovery via MML
,
1996,
ICML.
[4]
C Loehlin John,et al.
Latent variable models: an introduction to factor, path, and structural analysis
,
1986
.
[5]
John F. Lemmer,et al.
Causal Modeling
,
1993,
UAI.
[6]
M. Kendall,et al.
The Logic of Scientific Discovery.
,
1959
.
[7]
P. Spirtes,et al.
Causality From Probability
,
1989
.