em: EM step

In okamumu/Rsrat: A Software Reliability Assessment Tool with NHPP-Based Models

EM step

Description

Execute one EM step for software reliability models.

Usage

em_exp_emstep(params, data)

em_gamma_emstep(params, data, divide = 15L, eps = 1e-10)

em_llogis_emstep(params, data)

em_llogis_estep(params, data)

em_llogis_pllf(params, data, w1)

em_lnorm_emstep(params, data)

em_lxvmax_estep(params, data)

em_lxvmax_pllf(params, data, w1)

em_lxvmin_estep(params, data)

em_lxvmin_pllf(params, data, w1)

em_pareto_emstep(params, data)

em_tlogis_emstep_mo(params, data)

em_tlogis_estep(params, data)

em_tlogis_pllf(params, data, w0, w1)

em_tnorm_emstep(params, data)

em_txvmax_emstep_mo(params, data)

em_txvmax_estep(params, data)

em_txvmax_pllf(params, data, w0, w1)

em_txvmin_estep(params, data)

em_txvmin_pllf(params, data, w0, w1)

Arguments

params	A numeric vector for model parameters
data	A faultdata
divide	An integer for the number of integration points.
eps	A numeric value for tolerance error.
w1	A numeric value indicating the expected number of faults after te.
w0	A numeric value indicating the expected number of faults before 0.
...	Other parameters

Details

em_tlogis_emstep_mo is an emstep based on Marshall-Olkin-type (maximum) with Exp

em_tnorm_emstep has been modified by using EM for truncated distribution. Concretely, when the distribution is truncated at origin, the expected value is given by

\text{E}[h(X)|D] = \sum_{i=1}^k h(x_i) + \frac{k}{\overline{F}(0)} \int_{-\infty}^0 h(x) f(x) dx.

em_txvmax_emstep_mo is an emstep based on Marshall-Olkin-type (maximum) with Exp

Value

A list with an updated parameter vector (param), absolute difference of parameter vector (pdiff), log-likelihood function for a given parameter vector (llf), the number of total faults (total).

okamumu/Rsrat documentation built on Feb. 10, 2024, 11:07 p.m.