em: EM Algorithm (EM) Object
In wudongjie/fmmr6: The R Implementation of Finite Mixture Modellings on the R6 Class System
em R Documentation
EM Algorithm (EM) Object
Description
A concrete R6 object for EM Algorithm.
Value
Returns R6 object of class em.
Super classes
fmmr6::AbstractMethod -> fmmr6::EstimMethod -> em
Methods
Public methods
Inherited methods
Method fit()
Fit the model using the EM algorithm
Usage
em$fit(
algo = "em",
max_iter = 500,
start = "random",
rep = 1,
div_tol = 10,
verbose = F
)
Arguments
algo(character(1))
Specify the type of EM algorithm to use. Can choose em stands for the
conventional EM algorithm, cem stands for the classification EM algorithm,
and sem stands for the stochastic EM algorithm.
The default algorithm is em.
max_iter(numeric(1))
Specify the maximum number of iterations for the E-step-M-step loop.
The default number is 500.
start(character(1))
Specify the starting method of the EM algorithm.
Can either start from kmeans or random.
kmeans use the K-mean methods to put samples into latent classes.
random randomly assigns samples into latent classes.
The default method is kmeans.
rep(numeric(1))
Specify the number of reps EM-algorithm runs.
This parameter is designed for preventing the local maximum.
Each rep, the EM_algorithm generates a start.
It is only useful when start is random.
After all reps, the algorithm will pick the rep with maximum log likelihood.
The default value is 1
div_tol(numeric(1))
Divergence tolerence: the convergence process stops if there are div_tol number of
divergence ,i.e. the log-likelihood getting bigger.
verbose(bool(1))
Print the converging log-likelihood for all steps.
Method em_start()
Generate the start value for the EM algorithm.
Usage
em$em_start(start = "kmeans")
Arguments
start(character(1))
Specify the starting method of the EM algorithm.
Can either start from kmeans or random.
kmeans use the K-mean methods to put samples into latent classes.
random randomly assigns samples into latent classes.
The default method is kmeans.
Method em_algo()
Running the EM algorithm, given the start values, the data and number of classes.
Usage
em$em_algo(z, algo, max_iter, div_tol, verbose)
Arguments
z(matrix())
The matrix of the posterior probability.
algo(character(1))
Specify the type of EM algorithm to use. Can choose em stands for the
conventional EM algorithm, cem stands for the classification EM algorithm,
and sem stands for the stochastic EM algorithm.
The default algorithm is em.
max_iter(numeric(1))
Specify the maximum number of iterations for the E-step-M-step loop.
The default number is 500.
div_tol(numeric(1))
Divergence tolerence: the convergence process stops if there are div_tol number of
divergence ,i.e. the log-likelihood getting bigger.
verbose(bool(1))
Print the converging log-likelihood for all steps.
Method estep()
Given the data, start values, the number
of latent classes, compute the expected value of
the indicator variable.
Usage
em$estep(theta_update, pi_vector)
Arguments
theta_update(matrix())
Updated coefficients
pi_vector(matrix())
The matrix with the diagonal values representing the prior probability \pi.
Returns
return the probability matrix indicating the posterior probability of individual belonging to each class.
Method cstep()
Given the posterior probability, generate a matrix to assign
each individual to a class. The assignment based on which probability is the largest.
Usage
em$cstep(hidden)
Arguments
hidden(matrix())
The matrix of the posterior probability
Method sstep()
Given the posterior probability, generate a matrix to assign
each individual to a class. The assignment is randomly sampled based on the posterior probability.
Usage
em$sstep(hidden)
Arguments
hidden(matrix())
The matrix of the posterior probability
Method mstep()
Given the data, start values, the number
of latent classes, the indicator variable, the likelihood
function, solve the maximum value of the likelihood
function and update the parameter vectors theta.
Usage
em$mstep(hidden, theta, alpha = NULL, optim_method = "base")
Arguments
hidden(matrix())
The matrix of the posterior probability.
theta(matrix())
The matrix of the coefficients to fit.
optim_method(character(1))
The optimization method to use to fit the model.
The default is base.
Method clone()
The objects of this class are cloneable with this method.
Usage
em$clone(deep = FALSE)
Arguments
deepWhether to make a deep clone.
Author(s)
Dongjie Wu
wudongjie/fmmr6 documentation built on June 24, 2022, 2:48 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
| em | R Documentation |
EM Algorithm (EM) Object
Description
A concrete R6 object for EM Algorithm.
Value
Returns R6 object of class em.
Super classes
fmmr6::AbstractMethod -> fmmr6::EstimMethod -> em
Methods
Public methods
Inherited methods
Method fit()
Fit the model using the EM algorithm
Usage
em$fit( algo = "em", max_iter = 500, start = "random", rep = 1, div_tol = 10, verbose = F )
Arguments
algo(
character(1))
Specify the type of EM algorithm to use. Can chooseemstands for the conventional EM algorithm,cemstands for the classification EM algorithm, andsemstands for the stochastic EM algorithm. The default algorithm isem.max_iter(
numeric(1))
Specify the maximum number of iterations for the E-step-M-step loop. The default number is 500.start(
character(1))
Specify the starting method of the EM algorithm. Can either start fromkmeansorrandom.kmeansuse the K-mean methods to put samples into latent classes.randomrandomly assigns samples into latent classes. The default method iskmeans.rep(
numeric(1))
Specify the number of reps EM-algorithm runs. This parameter is designed for preventing the local maximum. Each rep, the EM_algorithm generates a start. It is only useful whenstartisrandom. After all reps, the algorithm will pick the rep with maximum log likelihood. The default value is 1div_tol(
numeric(1))
Divergence tolerence: the convergence process stops if there arediv_tolnumber of divergence ,i.e. the log-likelihood getting bigger.verbose(
bool(1))
Print the converging log-likelihood for all steps.
Method em_start()
Generate the start value for the EM algorithm.
Usage
em$em_start(start = "kmeans")
Arguments
start(
character(1))
Specify the starting method of the EM algorithm. Can either start fromkmeansorrandom.kmeansuse the K-mean methods to put samples into latent classes.randomrandomly assigns samples into latent classes. The default method iskmeans.
Method em_algo()
Running the EM algorithm, given the start values, the data and number of classes.
Usage
em$em_algo(z, algo, max_iter, div_tol, verbose)
Arguments
z(
matrix())
The matrix of the posterior probability.algo(
character(1))
Specify the type of EM algorithm to use. Can chooseemstands for the conventional EM algorithm,cemstands for the classification EM algorithm, andsemstands for the stochastic EM algorithm. The default algorithm isem.max_iter(
numeric(1))
Specify the maximum number of iterations for the E-step-M-step loop. The default number is 500.div_tol(
numeric(1))
Divergence tolerence: the convergence process stops if there arediv_tolnumber of divergence ,i.e. the log-likelihood getting bigger.verbose(
bool(1))
Print the converging log-likelihood for all steps.
Method estep()
Given the data, start values, the number of latent classes, compute the expected value of the indicator variable.
Usage
em$estep(theta_update, pi_vector)
Arguments
theta_update(
matrix())
Updated coefficientspi_vector(
matrix())
The matrix with the diagonal values representing the prior probability\pi.
Returns
return the probability matrix indicating the posterior probability of individual belonging to each class.
Method cstep()
Given the posterior probability, generate a matrix to assign each individual to a class. The assignment based on which probability is the largest.
Usage
em$cstep(hidden)
Arguments
hidden(
matrix())
The matrix of the posterior probability
Method sstep()
Given the posterior probability, generate a matrix to assign each individual to a class. The assignment is randomly sampled based on the posterior probability.
Usage
em$sstep(hidden)
Arguments
hidden(
matrix())
The matrix of the posterior probability
Method mstep()
Given the data, start values, the number of latent classes, the indicator variable, the likelihood function, solve the maximum value of the likelihood function and update the parameter vectors theta.
Usage
em$mstep(hidden, theta, alpha = NULL, optim_method = "base")
Arguments
hidden(
matrix())
The matrix of the posterior probability.theta(
matrix())
The matrix of the coefficients to fit.optim_method(
character(1))
The optimization method to use to fit the model. The default isbase.
Method clone()
The objects of this class are cloneable with this method.
Usage
em$clone(deep = FALSE)
Arguments
deepWhether to make a deep clone.
Author(s)
Dongjie Wu
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.