Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

code nutshell forest #1127

Open
wants to merge 3 commits into
base: master
Choose a base branch
from
Open

code nutshell forest #1127

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
90bd6a5
code nutshell forest
Aug 29, 2023
26ab65a
code nutshell forest fixed
Aug 29, 2023
70bcaf6
code nutshell forest changes
Sep 18, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
277 changes: 277 additions & 0 deletions slides/forests/rsrc/fig-nutshell-forest.R
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,277 @@
# PREREQ -----------------------------------------------------------------------

library(knitr)
library(mlr3)
library(mlr3learners)
library(mlr3viz)
library(ggplot2)
library(ISLR)
library(mlbench)
library(kernlab)
library(knitr)
library(gridExtra)
library(patchwork)


credit_data <- Credit
credit_data <- credit_data[, c("Rating", "Age","Income", "Limit", "Balance")]


options(digits = 3,
width = 65,
str = strOptions(strict.width = "cut", vec.len = 3))

# DATA -------------------------------------------------------------------------



set.seed(600000)
task = TaskRegr$new("credit_task",
backend = credit_data,
target = "Balance"
)
task$select(c("Limit"))


num_trees = c(1, 50, 50000)

learners = lapply(num_trees,
function (i) {lrn("regr.ranger",
num.trees = i)})




# PLOT 1 -----------------------------------------------------------------------

plot1 <- plot_learner_prediction(learners[[1]], task) +
ggtitle(paste0(num_trees[1], " Tree(s) for Credit Dataset")) +
guides(shape = FALSE, alpha = FALSE) + theme(axis.text=element_text(size=18),title=element_text(size = 22), axis.title=element_text(size=22),
legend.key.size = unit(1.5, 'cm'), legend.position = "none") +
xlab("Limit") + ylab("Balance")

ggsave(plot = plot1, "figure/nutshell_forest_ensemblesize_1.pdf", width = 8, height = 5.5)


# PLOT 2 -----------------------------------------------------------------------

plot2 <- plot_learner_prediction(learners[[2]], task) +
ggtitle(paste0(num_trees[2], " Tree(s) for Credit Dataset")) +
guides(shape = FALSE, alpha = FALSE) + theme(axis.text=element_text(size=18),title=element_text(size = 22), axis.title=element_text(size=22),
legend.key.size = unit(1.5, 'cm'), legend.position = "none") +
xlab("Limit") + ylab("Balance")

ggsave(plot = plot2, "figure/nutshell_forest_ensemblesize_2.pdf", width = 8, height = 5.5)


# PLOT 3 -----------------------------------------------------------------------

plot3 <- plot_learner_prediction(learners[[3]], task) +
ggtitle(paste0(num_trees[3], " Tree(s) for Credit Dataset")) +
guides(shape = FALSE, alpha = FALSE) + theme(axis.text=element_text(size=18),title=element_text(size = 22), axis.title=element_text(size=22),
legend.key.size = unit(1.5, 'cm'), legend.position = "none") +
xlab("Limit") + ylab("Balance")

ggsave(plot = plot3, "figure/nutshell_forest_ensemblesize_3.pdf", width = 8, height = 5.5)


# PLOT RF CLASSIFICATION

options(digits = 3,
width = 65,
str = strOptions(strict.width = "cut", vec.len = 3))

# DATA -------------------------------------------------------------------------
set.seed(6)
data = mlbench.twonorm(n = 100, d = 2)
data = as.data.frame(data)
X = as.matrix(data[, 1:2])
y = data$classes
colnames(data) <- c("x.1", "x.2", "Class")
data$Class <- as.numeric(data$Class)
data[data$Class == 2,3] <- 0
data$Class <- as.factor(data$Class)

task = TaskClassif$new("data_binary", data, target = "Class")
task$select(c("x.1", "x.2"))

# FUNCTIONS --------------------------------------------------------------------

num_trees = c(1, 10, 200)

learners_list = lapply(num_trees,
function (i) {lrn("classif.ranger",
num.trees = i, predict_type = "prob")})



# PLOT 1 CLASSIFICTAION -TREE 1 --------------------------------------------------------------------
set.seed(2)
features = task$feature_names
predict_sets = "train"
task$data(cols = features)
rr = mlr3::resample(task, learners_list[[1]]$clone(), mlr3::rsmp("holdout", ratio = 1), store_models = TRUE)
grid = mlr3viz:::predict_grid(rr$learners, rr$task, grid_points = 100L, expand_range = 0)
data = mlr3viz:::task_data(rr, predict_sets)
data$.predict_set <- as.numeric(data$.predict_set)


plot1 <- ggplot(grid, aes(
x = .data[[features[1L]]],
y = .data[[features[2L]]])
) +
geom_raster(aes(fill = .data[["response"]], alpha = .data[[".prob.response"]])) +
geom_point(
data = mlr3viz:::task_data(rr, predict_sets),size=3,
aes(shape = .data[["Class"]], fill = .data[["Class"]], color=.data[["Class"]]), color = "black") +
scale_fill_hue() +
scale_shape_manual(values = c(21,24)) +
scale_alpha_continuous(name = "Probability") +
labs(fill = "Class") +
guides(alpha = FALSE, color = guide_legend(
override.aes=list(shape = 19))) +
ggtitle(paste0("Tree 1")) +
theme(axis.text=element_text(size=14),title=element_text(size = 16), axis.title=element_text(size=14),
legend.key.size = unit(2.5, 'cm'), legend.title = element_text(size=14),
legend.text = element_text(size=14), legend.title.align=0.3, legend.position = "right")
ggsave(plot = plot1, "figure/nutshell_classif_forest_ensemblesize_1.pdf", width = 8, height = 5.5)

# PLOT 1 CLASSIFICATION - TREE 2 --------------------------------------------------------------------
set.seed(30)
features = task$feature_names
predict_sets = "train"
task$data(cols = features)
rr = mlr3::resample(task, learners_list[[1]]$clone(), mlr3::rsmp("holdout", ratio = 1), store_models = TRUE)
grid = mlr3viz:::predict_grid(rr$learners, rr$task, grid_points = 100L, expand_range = 0)
data = mlr3viz:::task_data(rr, predict_sets)
data$.predict_set <- as.numeric(data$.predict_set)

plot1_2 <- ggplot(grid, aes(
x = .data[[features[1L]]],
y = .data[[features[2L]]])
) +
geom_raster(aes(fill = .data[["response"]], alpha = .data[[".prob.response"]])) +
geom_point(
data = mlr3viz:::task_data(rr, predict_sets),size=3,
aes(shape = .data[["Class"]], fill = .data[["Class"]], color=.data[["Class"]]), color = "black") +
scale_fill_hue() +
scale_shape_manual(values = c(21,24)) +
scale_alpha_continuous(name = "Probability") +
labs(fill = "Class") +
guides(alpha = FALSE, color = guide_legend(
override.aes=list(shape = 19))) +
ggtitle(paste0("Tree 2")) +
theme(axis.text=element_text(size=14),title=element_text(size = 16), axis.title=element_text(size=14),
legend.key.size = unit(2.5, 'cm'), legend.title = element_text(size=14),
legend.text = element_text(size=14), legend.title.align=0.3, legend.position = "right")
ggsave(plot = plot1_2, "figure/nutshell_classif_forest_ensemblesize_1_TREE_2.pdf", width = 8, height = 5.5)

# PLOT 2 CLASSIFICATION
set.seed(2)
features = task$feature_names
predict_sets = "train"
task$data(cols = features)
rr = mlr3::resample(task, learners_list[[2]]$clone(), mlr3::rsmp("holdout", ratio = 1), store_models = TRUE)
grid = mlr3viz:::predict_grid(rr$learners, rr$task, grid_points = 100L, expand_range = 0)
data = mlr3viz:::task_data(rr, predict_sets)
data$.predict_set <- as.numeric(data$.predict_set)


plot2 <- ggplot(grid, aes(
x = .data[[features[1L]]],
y = .data[[features[2L]]])
) +
geom_raster(aes(fill = .data[["response"]], alpha = .data[[".prob.response"]])) +
geom_point(
data = mlr3viz:::task_data(rr, predict_sets),size=3,
aes(shape = .data[["Class"]], fill = .data[["Class"]], color=.data[["Class"]]), color = "black") +
scale_fill_hue() +
scale_shape_manual(values = c(21,24)) +
scale_alpha_continuous(name = "Probability") +
labs(fill = "Class") +
guides(alpha = FALSE, color = guide_legend(
override.aes=list(shape = 19))) +
ggtitle(paste0(num_trees[2], " Random Trees for Classification Task")) +
theme(axis.text=element_text(size=14),title=element_text(size = 16), axis.title=element_text(size=14),
legend.key.size = unit(1.5, 'cm'), legend.title = element_text(size=14),
legend.text = element_text(size=14), legend.title.align=0.3, legend.position = "right")

ggsave(plot = plot2, "figure/nutshell_classif_forest_ensemblesize_2.pdf", width = 8, height = 5.5)

# PLOT 3 CLASSIFICATION
set.seed(2)
features = task$feature_names
predict_sets = "train"
task$data(cols = features)
rr = mlr3::resample(task, learners_list[[3]]$clone(), mlr3::rsmp("holdout", ratio = 1), store_models = TRUE)
grid = mlr3viz:::predict_grid(rr$learners, rr$task, grid_points = 100L, expand_range = 0)
data = mlr3viz:::task_data(rr, predict_sets)
data$.predict_set <- as.numeric(data$.predict_set)



plot3 <- ggplot(grid, aes(
x = .data[[features[1L]]],
y = .data[[features[2L]]])
) +
geom_raster(aes(fill = .data[["response"]], alpha = .data[[".prob.response"]])) +
geom_point(
data = mlr3viz:::task_data(rr, predict_sets),size=3,
aes(shape = .data[["Class"]], fill = .data[["Class"]], color=.data[["Class"]]), color = "black") +
scale_fill_hue() +
scale_shape_manual(values = c(21,24)) +
scale_alpha_continuous(name = "Probability") +
labs(fill = "Class") +
guides(alpha = FALSE, color = guide_legend(
override.aes=list(shape = 19))) +
ggtitle(paste0("RF 1")) +
theme(axis.text=element_text(size=14),title=element_text(size = 16), axis.title=element_text(size=14),
legend.key.size = unit(2.5, 'cm'), legend.title = element_text(size=14),
legend.text = element_text(size=14), legend.title.align=0.3, legend.position = "right")

ggsave(plot = plot3, "figure/nutshell_classif_forest_ensemblesize_3.pdf", width = 8, height = 5.5)

# PLOT 3 CLASSIFICATION - TREE 2 --------------------------------------------------------------------

set.seed(30)
features = task$feature_names
predict_sets = "train"
task$data(cols = features)
rr = mlr3::resample(task, learners_list[[3]]$clone(), mlr3::rsmp("holdout", ratio = 1), store_models = TRUE)
grid = mlr3viz:::predict_grid(rr$learners, rr$task, grid_points = 100L, expand_range = 0)
data = mlr3viz:::task_data(rr, predict_sets)
data$.predict_set <- as.numeric(data$.predict_set)



plot3_2 <- ggplot(grid, aes(
x = .data[[features[1L]]],
y = .data[[features[2L]]])
) +
geom_raster(aes(fill = .data[["response"]], alpha = .data[[".prob.response"]])) +
geom_point(
data = mlr3viz:::task_data(rr, predict_sets),size=3,
aes(shape = .data[["Class"]], fill = .data[["Class"]], color=.data[["Class"]]), color = "black") +
scale_fill_hue() +
scale_shape_manual(values = c(21,24)) +
scale_alpha_continuous(name = "Probability") +
labs(fill = "Class") +
guides(alpha = FALSE, color = guide_legend(
override.aes=list(shape = 19))) +
ggtitle(paste0("RF 2")) +
theme(axis.text=element_text(size=14),title=element_text(size = 16), axis.title=element_text(size=14),
legend.key.size = unit(2.5, 'cm'), legend.title = element_text(size=14),
legend.text = element_text(size=14), legend.title.align=0.3, legend.position = "right")

ggsave(plot = plot3_2, "figure/nutshell_classif_forest_ensemblesize_3_TREE_3.pdf", width = 8, height = 5.5)


combined_plot_1 <- plot1 + plot1_2 + plot3 + plot3_2 + plot_layout(guides = "collect")
combined_plot_2 <- plot1_2 + plot3_2 + plot_layout(guides = "collect")

ggsave(plot = combined_plot_1, "figure/nutshell_classif_combined_1.pdf", width = 11, height = 10)
ggsave(plot = combined_plot_2, "figure/nutshell_classif_combined_2.pdf", width = 10, height = 4)


# PLOT K-NN BAGGED