MClik function

Likelihood Estimation for Markov Chains

Computes maximum likelihood estimates of transition probabilities for stationary Markov chain models, of order 0 (independence) to 3.

This is intended for use with Practical 6.1 of Davison (2003), not as production code.

MClik(d)

Arguments

• d: A sequence containing successive states of the chain

Returns

• order: order of fitted chain

• df: degrees of freedom using in fitting

• L: maximum log likelihood for each order

• AIC: Akaike information criterion for each order

• one: one-way marginal table of counts

• two: two-way margin table of transitions

• three: three-way marginal table of transitions

• four: four-way marginal table of transitions

References

Avery, P. J. and Henderson, D. A. (1999) Fitting Markov chain models to discrete state series such as DNA sequences. Applied Statistics, 48 , 53--61.

Davison, A. C. (2003) Statistical Models. Cambridge University Press. Section 6.1.

Author(s)

A. C. Davison (Anthony.Davison@epfl.ch)

Examples

data(intron)

fit <- MClik(intron)