Inference results for bloodtype.mln

Gibbs sampling

Command:
infer -i bloodtype.mln -e bloodtype.db -r bloodtype.results -q 'bloodtypeA(Linus),bloodtypeB(Linus),bloodtypeAB(Linus),bloodtypeN(Linus)' -p -maxSteps 20000 -burnMaxSteps -1

Output:
Done Gibbs sampling. 20000 samples per pred per chain (didn't converge at total of 200000 samples per pred)
Time taken for Gibbs sampling = 1.18 secs
total time taken = 1.24 secs

Results in bloodtype.results:
bloodtypeA(Linus) 0.440296
bloodtypeB(Linus) 0.181277
bloodtypeAB(Linus) 0.15475
bloodtypeN(Linus) 0.224448

Same command, but with changed order of query atoms:
infer -i bloodtype.mln -e bloodtype.db -r bloodtype.results -q 'bloodtypeB(Linus),bloodtypeA(Linus),bloodtypeAB(Linus),bloodtypeN(Linus)' -p -maxSteps 20000 -burnMaxSteps -1

Output:
Done Gibbs sampling. 20000 samples per pred per chain (didn't converge at total of 200000 samples per pred)
Time taken for Gibbs sampling = 1.1 secs
total time taken = 1.15 secs
Results in bloodtype.results:
bloodtypeB(Linus) 0.432097
bloodtypeA(Linus) 0.171873
bloodtypeAB(Linus) 0.108169
bloodtypeN(Linus) 0.288566

MC-SAT

Command (all default options):
infer -i bloodtype.mln -e bloodtype.db -r bloodtype.results -q 'bloodtypeA(Linus),bloodtypeB(Linus),bloodtypeAB(Linus),bloodtypeN(Linus)' -ms

Results in bloodtype.results (5 repetitions):
bloodtypeA(Linus) 0.186031
bloodtypeB(Linus) 0.363014
bloodtypeAB(Linus) 0.180032
bloodtypeN(Linus) 0.271023

bloodtypeA(Linus) 0.20003
bloodtypeB(Linus) 0.321018
bloodtypeAB(Linus) 0.170033
bloodtypeN(Linus) 0.309019

bloodtypeA(Linus) 0.257024
bloodtypeB(Linus) 0.222028
bloodtypeAB(Linus) 0.248025
bloodtypeN(Linus) 0.274023

bloodtypeA(Linus) 0.20103
bloodtypeB(Linus) 0.362014
bloodtypeAB(Linus) 0.174033
bloodtypeN(Linus) 0.263024

bloodtypeA(Linus) 0.19503
bloodtypeB(Linus) 0.322018
bloodtypeAB(Linus) 0.288021
bloodtypeN(Linus) 0.19503

Command (increased computational resources):
infer -i bloodtype.mln -e bloodtype.db -r bloodtype.results -q 'bloodtypeA(Linus),bloodtypeB(Linus),bloodtypeAB(Linus),bloodtypeN(Linus)' -ms -tries 10 -numStepsEveryMCSat 50 -burnMaxSteps -1

Results in bloodtype.results (3 repetitions):
bloodtypeA(Linus) 0.0060494
bloodtypeB(Linus) 0.789971
bloodtypeAB(Linus) 0.0520448
bloodtypeN(Linus) 0.152035

bloodtypeA(Linus) 0.0110489
bloodtypeB(Linus) 0.982952
bloodtypeAB(Linus) 0.0040496
bloodtypeN(Linus) 0.0020498

bloodtypeA(Linus) 0.0010499
bloodtypeB(Linus) 0.694981
bloodtypeAB(Linus) 0.20403
bloodtypeN(Linus) 0.10004

Command (as previous, changed order of query atoms):
infer -i bloodtype.mln -e bloodtype.db -r bloodtype.results -q 'bloodtypeB(Linus),bloodtypeA(Linus),bloodtypeAB(Linus),bloodtypeN(Linus)' -ms -tries 10 -numStepsEveryMCSat 50 -burnMaxSteps -1

Results in bloodtype.results (3 repetitions):
bloodtypeB(Linus) 0.811969
bloodtypeA(Linus) 0.175032
bloodtypeAB(Linus) 0.0070493
bloodtypeN(Linus) 0.0060494

bloodtypeB(Linus) 0.0040496
bloodtypeA(Linus) 0.769973
bloodtypeAB(Linus) 0.174033
bloodtypeN(Linus) 0.0530447

bloodtypeB(Linus) 0.0020498
bloodtypeA(Linus) 0.333017
bloodtypeAB(Linus) 0.436006
bloodtypeN(Linus) 0.236026

Command (only one query atom):
infer -i bloodtype.mln -e bloodtype.db -r bloodtype.results -q 'bloodtypeA(Linus)' -ms -tries 10 -numStepsEveryMCSat 50 -burnMaxSteps -1

Results in bloodtype.results (3 repetitions):
bloodtypeA(Linus) 0.99995

bloodtypeA(Linus) 0.99995

bloodtypeA(Linus) 0.99895

Simulated tempering

Summary: similar observations as for previous two: dependence on order/number of query atoms.