mmpp.obsolete          package:HiddenMarkov          R Documentation

_M_a_r_k_o_v _M_o_d_u_l_a_t_e_d _P_o_i_s_s_o_n _P_r_o_c_e_s_s - _O_b_s_o_l_e_t_e _F_u_n_c_t_i_o_n_s

_D_e_s_c_r_i_p_t_i_o_n:

     These functions are obsolete and will ultimately be removed from
     the package. Please change to the revised versions: 'BaumWelch',
     'Estep.mmpp', 'forwardback.mmpp', 'simulate' or 'logLik'.

_U_s_a_g_e:

     backward0.mmpp(tau, Q, lambda)
     forward0.mmpp(tau, Q, delta, lambda)

     logLikmmpp(tau, Q, delta, lambda)

     Estep0.mmpp(tau, Q, delta, lambda)

     Baum.Welch.mmpp(tau, Q, delta, lambda, nonstat = TRUE,
                     maxiter = 500, tol = 1e-05, prt = TRUE,
                     converge = expression(diff < tol))
     Baum.Welch0.mmpp(tau, Q, delta, lambda, nonstat = TRUE,
                      maxiter = 500, tol = 1e-05, prt = TRUE,
                      converge = expression(diff < tol))

     sim.mmpp(n, initial, Q, lambda)

_A_r_g_u_m_e_n_t_s:

     tau: vector containing the interevent times. Note that the first
          event is at time zero.

       Q: the infinitesimal generator matrix of the Markov process.

  lambda: a vector containing the Poisson rates.

   delta: is the marginal probability distribution of the m hidden
          states at time zero.

       n: number of Poisson events to be simulated.

 initial: integer, being the initial hidden Markov state (1, cdots, m).

 nonstat: is logical, 'TRUE' if the homogeneous Markov chain is assumed
          to be non-stationary, default.

 maxiter: is the maximum number of iterations, default is 500.

     tol: is the convergence criterion, being the difference between
          successive values of the log-likelihood; default is 0.00001.

     prt: is logical, and determines whether information is printed at
          each iteration; default is 'TRUE'.

converge: is an expression giving the convergence criterion.

_D_e_t_a_i_l_s:

     The functions with a suffix of zero are non-scaled, and hence will
     have numerical problems for series containing larger numbers of
     events; and are _much_ slower.

     These functions use the algorithm given by Ryden (1996) based on
     eigenvalue decompositions.

