MARKOV STUDY P,L where P is a matrix file of transition probabilities determines the class structure of the Markov chain and gives the results from the edit line L onwards as shown in the following example: MATRIX P /// T1 T2 T3 T4 T5 T1 0.4 0 0 0 0.6 T2 0.9 0 0 0.1 0 T3 0 0 0.2 0.8 0 T4 0 0 0.8 0.2 0 T5 0.7 0 0 0 0.3 MAT SAVE P MARKOV STUDY P,CUR+1 Structure of Markov chain P of 5 states: Class structure saved in matrix file MCLASS.M 2 recurrent classes of states: 1 (2): T1 T5 2 (2): T3 T4 1 transient state: T2 By default the results are obtained by finding the transitive closure of the digraph determined by P. By using the specification SVD=1 the same task is accomplished by computing the singular value decomposition of I-P. Then also the steady-state probabilities for each recurrent classes are calculated and given as the second column of matrix MCLASS.M In the above example this gives LOADM MCLASS.M,(C7),CUR+1 Class_structure_of_P_(Transient_states=0) Class Prob T1 1 0.53846 T2 0 0.00000 T3 2 0.50000 T4 2 0.50000 T5 1 0.46154 M = More information on Markov chains

More information on Survo from www.survo.fi

Copyright © Survo Systems 2001-2012.

webmaster'at'survo.fi

Copyright © Survo Systems 2001-2012.

webmaster'at'survo.fi