Version numbers follow Semantic Versioning.
2020-03-18
2020-03-02
straingsAsFactors update2019-12-22
sim_TS2019-07-28
2019-07-24
print/cat has been replaced with message messages.duntrun{} with donttest{})messageq replaces qprint2019-07-10
2019-07-09
2019-07-09
LDA_TS function level, the separate inputs for data tables (document_term_table and document_covariate_table) have been merged into a single input data, which can be just the document_term_table or a list including the document_term_table and optionally also a document_covariate_table. If covariates aren’t provided, the function now constructs a covariate table assuming equi-spaced observations. If using a list, the function assumes that one and only one element of the list will have a name containing the letters “term”, and at most one element containing the letters “covariate” (regular expressions are used for matching). (addresses issue 119)timename has been moved from within the TS_controls_list to a main argument in all associated functions.LDA_controls_list, TS_controls_list, or LDA_TS_controls_list inputs now take general list inputs (so LDA_TS does not need to have a nested set of control functions). Each control list is passed through a function (LDA_set_control, TS_control, or LDA_TS_control) to set any non-input values to their defaults. This also allows the removal of those controls list class definitions. (addresses issue 130)control input in the plot call in the example in the README (addresses issue 116)?logLik.LDA_VEM for references.devAskNewPage to help flip through multiple outputs, but were only resetting it with devAskNewPage(FALSE) at the end of a clean execution. The code has been updated with on.exit(devAskNewPage(FALSE)), which accounts for failed executions. (addresses issue 118)summarize_TS has been renamed package_TS to align with the other package_ functions that package model output.sim_LDA_data simulates an LDA model’s document-term-matrixsim_TS_data simulates an TS model’s document-topic distribution matrixsim_LDA_TS_data simulates an LDA_TS model’s document-term-matrixsoftmax and logsumexp are added as utility functionsTSTS was named “deviance”. The output has been updated to return the AIC.AIC method with logLik method for TS_fit2019-02-11
AIC.LDA_VEM() now uses the number of parameters as reported from logLik to calculate AIC.document_weights() function is provided to allow for appropriate weighting of documents by their sizes (number of words) so that an average-length document is 1.weights = NULL.multinom_TS() and multinom_TS_chunk() now is optional via memoise_fun() and is controlled through the TS controls list.LDA_set(), LDA_TS(), and TS() now all have default plotting options on their outputs.plot.TS() provides MCMC diagnostic plots and summary plots.plot.LDA_TS() plots produce the combination of plots.data(rodents).The comparison vignette provides a step-by-step comparison of the LDATS pipeline to the analysis in Christensen et al. 2018.
The key differences are as follows:
* The `document_term_table` in Christensen *et al.* 2018 was adjusted to account for variable trapping effort. The data included in LDATS are not adjusted, so that sampling periods can be weighted appropriately.
* The LDA model selection criterion has changed (see LDA model AIC calculation, above), so that LDATS now identifies 6 topics compared to the 4 topics found in the paper.
* LDATS will by default weight sampling periods according to the number of terms (see Document weighting, above).
* Despite these changes, the updated LDATS pipeline gives qualitatively similar results to the analysis in Christensen *et al.* 2018. 2017-11-16