run.e0.mcmc: Running Bayesian Hierarchical Model for Life Expectancy via...

View source: R/run_mcmc.R

run.e0.mcmcR Documentation

Running Bayesian Hierarchical Model for Life Expectancy via Markov Chain Monte Carlo

Description

Runs (or continues running) MCMCs for simulating the life expectancy for all countries of the world, using a Bayesian hierarchical model.

Usage

run.e0.mcmc(sex = c("Female", "Male"), nr.chains = 3, iter = 160000, 
    output.dir = file.path(getwd(), "bayesLife.output"), 
    thin = 10, replace.output = FALSE, annual = FALSE,
    start.year = 1873, present.year = 2020, wpp.year = 2019, 
    my.e0.file = NULL, my.locations.file = NULL, constant.variance = FALSE, 
    seed = NULL, parallel = FALSE, nr.nodes = nr.chains, compression.type = 'None',
    verbose = FALSE, verbose.iter = 100, mcmc.options = NULL, ...)
    
continue.e0.mcmc(iter, chain.ids = NULL, 
    output.dir = file.path(getwd(), "bayesLife.output"), 
    parallel = FALSE, nr.nodes = NULL, auto.conf = NULL, 
    verbose = FALSE, verbose.iter = 10, ...)

Arguments

sex

Sex for which to run the simulation.

nr.chains

Number of MCMC chains to run.

iter

Number of iterations to run in each chain. In addition to a single value, it can have the value ‘auto’ for an automatic assessment of the convergence. In such a case, the function runs for the number of iterations given in the global option auto.conf list (see e0mcmc.options), then checks if the MCMCs converged (using the auto.conf settings). If it did not converge, the procedure is repeated until convergence is reached or the number of repetition exceeded auto.conf$max.loops.

output.dir

Directory which the simulation output should be written into.

thin

Thinning interval between consecutive observations to be stored on disk.

replace.output

If TRUE, existing outputs in output.dir will be replaced by results of this simulation.

annual

If TRUE, the model will be trained based on annual data. in such a case, argument my.e0.file must be used to provide the annual observed data.

start.year

Start year for using historical data.

present.year

End year for using historical data.

wpp.year

Year for which WPP data is used. The functions loads a package called wppx where x is the wpp.year and uses the e0* datasets.

my.e0.file

File name containing user-specified e0 time series for one or more countries. See Details below.

my.locations.file

File name containing user-specified locations. See Details below.

constant.variance

Logical indicating if the model should be estimated using constant variance. It should only be used if the standard deviation lowess is to be analysed, see compute.loess.

seed

Seed of the random number generator. If NULL no seed is set. It can be used to generate reproducible results.

parallel

Logical determining if the simulation should run multiple chains in parallel. If it is TRUE, the package snowFT is required.

nr.nodes

Relevant only if parallel is TRUE. It gives the number of nodes for running the simulation in parallel. By default it equals to the number of chains.

compression.type

One of ‘None’, ‘gz’, ‘xz’, ‘bz’, determining type of a compression of the MCMC files.

verbose

Logical switching log messages on and off.

verbose.iter

Integer determining how often (in number of iterations) log messages are outputted during the estimation.

mcmc.options

List of options that overwrites global MCMC options as defined in e0mcmc.options. Type e0mcmc.options() to view default values.

auto.conf

In continue.e0.mcmc, one can overwrite the global auto.conf option, see e0mcmc.options for its definition. This argument is only used if the function argument iter is set to ‘auto’.

...

Additional parameters to be passed to the function snowFT::performParallel, if parallel is TRUE.

chain.ids

Array of chain identifiers that should be resumed. If it is NULL, all existing chains in output.dir are resumed.

Details

The function run.e0.mcmc uses a set of global options (for priors, initial values etc.), possibly modified by the mcmc.options argument. One can also modify these options using e0mcmc.options. Call e0mcmc.options() for the full set of options. Function continue.e0.mcmc inherits its set of options from the corresponding run.e0.mcmc call.

The function run.e0.mcmc creates an object of class bayesLife.mcmc.meta and stores it in output.dir. It launches nr.chains MCMCs, either sequentially or in parallel. Parameter traces of each chain are stored as (possibly compressed) ASCII files in a subdirectory of output.dir, called mcx where x is the identifier of that chain. There is one file per parameter, named after the parameter with the suffix “.txt”, possibly followed by a compression suffix if compression.type is given. Country-specific parameters have the suffix _countryc where c is the country code. In addition to the trace files, each mcx directory contains the object bayesLife.mcmc in binary format. All chain-specific files are written into disk after the first, last and each i-th (thinned) iteration, where i is given by the global option buffer.size.

Using the function continue.e0.mcmc one can continue simulating an existing MCMCs by iter iterations for either all or selected chains. The global options used for generating the existing MCMCs will be used. Only the auto.conf option can be overwritten by passing the new value as an argument.

The function loads observed data (further denoted as WPP dataset), depending on the specified sex, from the e0F (e0M) and e0F_supplemental (e0M_supplemental) datasets in a wppx package where x is the wpp.year. It is then merged with the include dataset that corresponds to the same wpp.year. The argument my.e0.file can be used to overwrite those default data. Such a file can include a subset of countries contained in the WPP dataset, as well as a set of new countries. In the former case, the function replaces the corresponding country data from the WPP dataset with values in this file. Only columns are replaced that match column names of the WPP dataset, and in addition, columns ‘last.observed’ and ‘include_code’ are used, if present. Countries are merged with WPP using the column ‘country_code’. In addition, in order the countries to be included in the simulation, in both cases (whether they are included in the WPP dataset or not), they must be contained in the table of locations (UNlocations). In addition, their corresponding ‘include_code’ must be set to 2. If the column ‘include_code’ is present in my.e0.file, its value overwrites the default include code, unless is -1.

If annual is TRUE the default WPP dataset is not used and the my.e0.file argument must provide the dataset to be used for estimation. Its time-related columns should be single years.

The default UN table of locations mentioned above can be overwritten/extended by using a file passed as the my.locations.file argument. Such a file must have the same structure as the UNlocations dataset. Entries in this file will overwrite corresponding entries in UNlocations matched by the column ‘country_code’. If there is no such entry in the default dataset, it will be appended. This option of appending new locations is especially useful in cases when my.e0.file contains new countries/regions that are not included in UNlocations. In such a case, one must provide a my.locations.file with a definition of those countries/regions.

For simulation of the hyperparameters of the Bayesian hierarchical model, all countries are used that are included in the WPP dataset, possibly complemented by the my.e0.file, that have include_code equal to 2. The hyperparameters are used to simulate country-specific parameters, which is done for all countries with include_code equal 1 or 2. The following values of include_code in my.e0.file are recognized: -1 (do not overwrite the default include code), 0 (ignore), 1 (include in prediction but not estimation), 2 (include in both, estimation and prediction). Thus, the set of countries included in the estimation and prediction can be fully specified by the user.

Optionally, my.e0.file can contain a column called last.observed containing the year of the last observation for each country. In such a case, the code would ignore any data after that time point. Furthermore, the function e0.predict fills in the missing values using the median of the BHM procedure (stored in e0.matrix.reconstructed of the bayesLife.prediction object). For last.observed values that are below a middle year of a time interval [t_i, t_{i+1}] (computed as t_i+3) the last valid data point is the time interval [t_{i-1}, t_i], whereas for values larger equal a middle year, the data point in [t_i, t_{i+1}] is valid.

The package contains a dataset called ‘my_e0_template’ (in ‘extdata’ directory) which is a template for user-specified my.e0.file.

Value

An object of class bayesLife.mcmc.set which is a list with two components:

meta

An object of class bayesLife.mcmc.meta.

mcmc.list

A list of objects of class bayesLife.mcmc, one for each MCMC.

Author(s)

Hana Sevcikova, Patrick Gerland contributed to the documentation.

References

J. L. Chunn, A. E. Raftery, P. Gerland, H. Sevcikova (2013): Bayesian Probabilistic Projections of Life Expectancy for All Countries. Demography 50(3):777-801. <doi:10.1007/s13524-012-0193-x>

See Also

get.e0.mcmc, summary.bayesLife.mcmc.set, e0mcmc.options, e0.predict.

Examples

## Not run: 
m <- run.e0.mcmc(nr.chains = 1, iter = 5, thin = 1, verbose = TRUE)
summary(m)
m <- continue.e0.mcmc(iter = 5, verbose = TRUE)
summary(m)
## End(Not run)

bayesLife documentation built on Sept. 16, 2023, 9:07 a.m.