BNW using Octave instead of Matlab.

This version of BNW should perform the same as the original version. The only difference is that it uses Octave instead of Matlab when running BayesNet Toolbox during parameter learning.

I am calling this BNW_1.02. It can be accessed at:
compbio.uthsc.edu/BNW_1.02

1 files changed, 26 insertions, 0 deletions

diff --git a/sourcecodes/bnt-master/KPMstats/mc_stat_distrib.m b/sourcecodes/bnt-master/KPMstats/mc_stat_distrib.m
new file mode 100644
index 00000000..a5806092
--- /dev/null
+++ b/sourcecodes/bnt-master/KPMstats/mc_stat_distrib.m
@@ -0,0 +1,26 @@
+function pi = mc_stat_distrib(P)
+% MC_STAT_DISTRIB Compute stationary distribution of a Markov chain
+% function pi = mc_stat_distrib(P)
+% 
+% Each row of P should sum to one; pi is a column vector
+
+% Kevin Murphy, 16 Feb 2003
+
+% The stationary distribution pi satisfies pi P = pi
+% subject to sum_i pi(i) = 1,  0 <= pi(i) <= 1
+% Hence
+% (P'  0n   (pi  = (pi 
+%  1n  0)    1)     1)
+% or P2 pi2 = pi2.
+% Naively we can solve this using (P2 - I(n+1)) pi2 = 0(n+1)
+% or P3 pi2 = 0(n+1), i.e., pi2 = P3 \ zeros(n+1,1)
+% but this is singular (because of the sum-to-one constraint).
+% Hence we replace the last row of P' with 1s instead of appending ones to create P2, 
+% and similarly for pi.
+
+n = length(P);
+P4 = P'-eye(n);
+P4(end,:) = 1;
+pi = P4 \ [zeros(n-1,1);1];
+
+