n8thangreen / treesimr

:palm_tree::left_right_arrow::pill: Easy, robust forward simulating decision trees, calculating cost-effectiveness and PSA.

treeSimR

An R package for easy, robust forward simulating probability decision trees, calculating cost-effectiveness and probability sensitivity analysis (PSA).

Introduction

The package currently contains functions to:

• read-in and check tree object
• simulate final expected outcomes
• Monte-Carlo simulate multiple simulations

TODO

• iteratively collapse expected outcome (from right to left)
• iteratively collapse chance nodes (from right to left)
• optimal decision function (iterative from right to left)

The package leans heavily on the data.tree package, (introduction here and examples here ).

Installing treeSimR

To install the development version from github:

library(remotes)
install_github("n8thangreen/treeSimR")

Then, to load the package, use:

library("treeSimR")

See package Vignettes for more details and examples.

For plotting functions (e.g. C-E plane, C-E curve, risk profile (with uncertainty), tornado, spider, …) see the plotCostEffectiveness package.

License

GPL-3 ©

treesimr's Issues

write head(osNode)/tail Method

write wrapper for sampled tree using vignette

chance node reduction

part of solving a tree

add testthat tests

fillin missing tree probs using data.tree operations

remove all tb references

duplicate previous IDEA and LTBI code output

link output with plotting functions

plotting functions

start with previous IDEA plots
#8

tornado plots
spider graphs

strategy region plots

write function to do PSA

decision reduction

from right to left

deterministic sensitivity analysis

one-way and two-way

binomial (prior) branching probability sensitivity analysis sampling

plot decision trees with H2O