statsmodels.tsa.regime_switching.markov_regression.MarkovRegression.regime_transition_matrix¶

MarkovRegression.regime_transition_matrix(params, exog_tvtp=None)¶

Construct the left-stochastic transition matrix

Notes

This matrix will either be shaped (k_regimes, k_regimes, 1) or if there are time-varying transition probabilities, it will be shaped (k_regimes, k_regimes, nobs).

The (i,j)th element of this matrix is the probability of transitioning from regime j to regime i; thus the previous regime is represented in a column and the next regime is represented by a row.

It is left-stochastic, meaning that each column sums to one (because it is certain that from one regime (j) you will transition to some other regime).