AnnealedSMC is a recently proposed algorithm (annealed sequential Monte Carlo) with the adaptive determination of annealing parameters based on the relative conditional effective sample size for Bayesian phylogenetics. This algorithm provides an unbiased estimator for the probability of the data. This unbiasedness property can be used for the purpose of testing the correctness of posterior simulation software. Unlike the previous SMC methods in phylogenetics, the annealed SMC has the same state space for all the intermediate distributions, which allows standard Markov chain Monte Carlo (MCMC) tree moves to be utilized as the basis for SMC proposal distributions.
There are several options available to install the package:
Simply add the following lines (replacing 1.0.0 by the current version (see git tags)):
repositories {
mavenCentral()
jcenter()
maven {
url "http://people.stat.sfu.ca/~lwa68/maven/"
}
}
dependencies {
compile group: 'ca.sfu.stat', name: 'annealedSMC', version: '1.0.0'
}
- Check out the source
git clone git@github.com:liangliangwangsfu/annealedSMC.git
- Compile using
./gradlew installDist
- Add the jars in
build/install/annealedSMC/lib/
into your classpath
- Check out the source
git clone git@github.com:liangliangwangsfu/annealedSMC.git
- Type
gradle eclipse
from the root of the repository - From eclipse:
Import
inFile
menuImport existing projects into workspace
- Select the root
- Deselect
Copy projects into workspace
to avoid having duplicates
-useDataGenerator true -nThousandIters 0.001 -nTax 15 -len 200 -sequenceType DNA -generateDNAdata true -useDataGen4GTRGammaI false -nThreads 2 -treeRate 10 -deltaProposalRate 10 -useNonclock true -useSlightNonclock false -sdScale 0.3 -iterScalings 100 -methods ANNEALING -resamplingStrategy ESS -nAnnealing 10000 -nSubsampling 10000 -alphaSMCSampler 0.9999 -nSitesPerIndex 10 -essRatioThreshold 0.5 -adaptiveTempDiff true -runDSMCusingadaptiveTemp false -adaptiveType 0 -csmc_trans2tranv 2.0 -mb_trans2tranv 2.0 -setJC false -fixtratioInMb true -treePrior unconstrained:exp(10) -fixNucleotideFreq true -nReplica 1 -repPerDataPt 10 -mainRand 452 -gen.rand 345 -useCESS true -useNNI true -useLIS false -usenewSS false -useRevSS false -ntempSS 50 -mcmcfac 1 -usenewDSMC true -mrBayesPath /Users/oudomame/Dropbox/phyloSoftware/mrbayes-3.2.6/src//mb -neighborPath /Users/oudomame/Dropbox/phyloSoftware/phylip-3.69/exe//neighbor
For settings in our experimental results, please refers to ``https://github.com/shijiaw/annealingExperiment''.