README.md

In MateusMaiaDS/gpbart: GP-BART: Gaussian Processes Bayesian Additive Regression Trees

GP-BART

Installation

You can install the development version:

devtools::install_github("MateusMaiaDS/gpbart")

Setting a example

This a vignette to explain how to run a simple example of the model, setting its own prior and its hyperparameters. To start we going to use the friedman example as the dataset to be used.

```{r setup, eval = FALSE} library(gpbart) library(mlbench)

Setting simulation parameters

n <- 100 seed <- 42 sd <- 0.1

Loading the data

fried_data <- as.data.frame(mlbench::mlbench.friedman1(n = n,sd = 0.1))

Setting a cross-validation object

cv_obj <- k_fold(data = fried_data,dependent_variable = "y", k_partitions = 5,seed = seed,as_data_frame = TRUE)

Selecting one fold

fold <- 2 x_train <- cv_obj[[fold]]$x_train y_train <- c(unlist(cv_obj[[fold]]$y_train)) x_test <- cv_obj[[fold]]$x_test y_test <- cv_obj[[fold]]$y_test

### Running the model

To run the model we would have:

```{r, eval = FALSE}
gp_bart_mod <- gpbart(x_train = x_train,
                       y = c(unlist(y_train)),
                       x_test = x_test,
                       n_tree = 20)

MateusMaiaDS/gpbart documentation built on Jan. 26, 2024, 7:07 a.m.