Changelog

Source: NEWS.md

rTFT 0.1.1

Bug fixes

• From version 2.10.0, loo::loo_compare() gives the compared models in a column rather than as row names. A test and the vignette “Comparing and averaging models” read the row names; both now read the models from either place. The functions of the package are unchanged.

Other changes

• The package has a DOI, given by citation("rTFT") and the README.
• The package is checked on Linux, macOS and Windows with each push.
• The package can be installed from R-universe.
• The website has a page title and description for search engines.

rTFT 0.1.0

First release.

Fitting

• build_hazard_data() builds the table of intervals at risk from failure times that are known-time, interval-censored or a mixture of both, under any temperature trajectory. ramp_path() builds the trajectory of a linear ramp, and assess_ramp_linearity() compares a logged ramp with a straight line.
• fit_hazard() fits the model using Bayesian inference in Stan through brms, with a choice of functions for the failure rate (M_catalogue) and for the baseline (h0_catalogue). The failure rate may depend on covariates and on group-level effects.
• costressor() declares a co-stressor (e.g., dissolved oxygen), with its main effect and, when asked for, its interaction with temperature. A model may have more than one co-stressor.
• build_hazard_prior() shows and sets the priors.

Checking a fitted model

• check_survival() and plot_check() compare the proportion of individuals observed to survive with the proportions simulated from the model.
• hazard_residuals() and plot_residuals() give one randomised residual per individual.

Derived quantities

• derive_tdt_ttl() derives the static critical thermal maximum, the thermal death time curves and the thermal tolerance landscape, and derive_dctmax() the dynamic critical thermal maximum under a ramp, each with its posterior uncertainty.
• plot_tdt(), plot_ttl() and plot_ttl_3d() draw them.
• M_draws() returns the posterior draws of the failure rate across temperatures.

Comparing and averaging models

• loo_hazard() and kfold_hazard() compare models by cross-validation that leaves out whole individuals or trials.
• hazard_weights() gives the weights of the models, and average_tdt_ttl(), average_dctmax() and average_survival() average the derived quantities and the predictions.

Predicting survival

• predict_survival() predicts survival along a record of temperature (and of any co-stressors recorded with it), and reports how much of the record and of the predicted risk lies outside the temperatures and exposure times of the assay. plot_survival() and plot_risk() draw the prediction.

Data

• static_assay, ramp_assay and stream_record are simulated datasets used in the help pages and the vignettes.