Simulations/Simulation_TimeComparison_cluster.R
In tulliapadellini/energytree: Conditional Inference trees for mixed type data
# Inizialization ----------------------------------------------------------
# Loading the libraries
library(fda.usc)
library(roahd)
library(energy)
library(entropy)
library(partykit)
library(igraph)
library(NetworkDistance)
library(microbenchmark)
# Loading the dataset
load("Regression_Sim_Dataset.RData")
# Datasets ----------------------------------------------------------------
norm01 <- function(x){(x-min(x))/(max(x)-min(x))}
### DATASET 1 (20 observations) ###
# Response
resp1 <- lapply(data, function(x) x$Y[1:20])[[1]]
# Building the numerical variable
foo <- fda.usc::min.basis(lapply(data, function(x) x$V1)[[2]], numbasis = 15)
fd3 <- fda.usc::fdata2fd(foo$fdata.est,
type.basis = "bspline",
nbasis = foo$numbasis.opt)
foo$coef <- t(fd3$coefs)
# Building the graph variable
conn.prob <- norm01(resp1)
conn.prob[which(conn.prob < 0.05)] <- conn.prob[which(conn.prob < 0.05)] + 0.1
set.seed(12345)
graph.list <- lapply(conn.prob, function(p){sample_gnp(100, p)})
# Covariates for the full mixed model
cov.list1 <- list(lapply(data, function(x) x$V1[1:20])[[1]], graph.list, foo$coef[1:20,15])
### DATASET 2 (100 observations) ###
# Response
resp2 <- lapply(data, function(x) x$Y[1:100])[[1]]
# Building the numerical variable
foo <- fda.usc::min.basis(lapply(data, function(x) x$V1)[[2]], numbasis = 15)
fd3 <- fda.usc::fdata2fd(foo$fdata.est,
type.basis = "bspline",
nbasis = foo$numbasis.opt)
foo$coef <- t(fd3$coefs)
# Building the graph variable
conn.prob <- norm01(resp2)
conn.prob[which(conn.prob < 0.05)] <- conn.prob[which(conn.prob < 0.05)] + 0.1
set.seed(12345)
graph.list <- lapply(conn.prob, function(p){sample_gnp(100, p)})
# Covariates for the full mixed model
cov.list2 <- list(lapply(data, function(x) x$V1[1:100])[[1]], graph.list, foo$coef[1:100,15])
### DATASET 3 (200 observations) ###
# Response
resp3 <- lapply(data, function(x) x$Y)[[1]]
# Building the numerical variable
foo <- fda.usc::min.basis(lapply(data, function(x) x$V1)[[2]], numbasis = 15)
fd3 <- fda.usc::fdata2fd(foo$fdata.est,
type.basis = "bspline",
nbasis = foo$numbasis.opt)
foo$coef <- t(fd3$coefs)
# Building the graph variable
conn.prob <- norm01(resp3)
conn.prob[which(conn.prob < 0.05)] <- conn.prob[which(conn.prob < 0.05)] + 0.1
set.seed(12345)
graph.list <- lapply(conn.prob, function(p){sample_gnp(100, p)})
# Covariates for the full mixed model
cov.list3 <- list(lapply(data, function(x) x$V1)[[1]], graph.list, foo$coef[,15])
# functions ---------------------------------------------------------------
## distances & expansions inside
source("functions.R")
mb_cl_fun_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_fun_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_fun_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# fun_test1 ---------------------------------------------------------------
## distances inside, expansions outside
source("fun_test1.R")
mb_cl_ft1_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft1_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft1_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# fun_test2 ---------------------------------------------------------------
## variable selection distances & expansions outside
source("fun_test2.R")
mb_cl_ft2_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft2_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft2_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 10,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# fun_test3 ---------------------------------------------------------------
## (variable selection + clustering) distances & expansions outside
source("fun_test3.R")
mb_cl_ft3_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft3_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft3_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 10,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# fun_test4 ---------------------------------------------------------------
## (variable selection + clustering + expansions) distances & expansions outside
source("fun_test4.R")
mb_cl_ft4_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft4_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft4_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 10,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# Comparison --------------------------------------------------------------
et_cl <- rbind(mb_cl_fun_1$time,
mb_cl_fun_2$time,
mb_cl_fun_3$time,
mb_cl_ft1_1$time,
mb_cl_ft1_2$time,
mb_cl_ft1_3$time,
mb_cl_ft2_1$time,
mb_cl_ft2_2$time,
mb_cl_ft2_3$time,
mb_cl_ft3_1$time,
mb_cl_ft3_2$time,
mb_cl_ft3_3$time,
mb_cl_ft4_1$time,
mb_cl_ft4_2$time,
mb_cl_ft4_3$time)
et_cl <- et_cl/1000000000
summary(et_cl)
save(mb_cl_fun_1,
mb_cl_fun_2,
mb_cl_fun_3,
mb_cl_ft1_1,
mb_cl_ft1_2,
mb_cl_ft1_3,
mb_cl_ft2_1,
mb_cl_ft2_2,
mb_cl_ft2_3,
mb_cl_ft3_1,
mb_cl_ft3_2,
mb_cl_ft3_3,
mb_cl_ft4_1,
mb_cl_ft4_2,
mb_cl_ft4_3,
file = 'TimeComparison_cluster.RData')
### Legend ###
# cl: cluster
# fun: distances & expansions inside
# ft1: distances inside, expansions outside
# ft2: variable selection distances & expansions outside
# ft3: (variable selection + clustering) distances & expansions outside
# ft4: (variable selection + clustering + expansions) distances & expansions outside
# _1: small dataset (20 observations)
# _2: medium dataset (100 observations)
# _3: large dataset (200 observations)
tulliapadellini/energytree documentation built on May 14, 2020, 8:06 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
# Inizialization ----------------------------------------------------------
# Loading the libraries
library(fda.usc)
library(roahd)
library(energy)
library(entropy)
library(partykit)
library(igraph)
library(NetworkDistance)
library(microbenchmark)
# Loading the dataset
load("Regression_Sim_Dataset.RData")
# Datasets ----------------------------------------------------------------
norm01 <- function(x){(x-min(x))/(max(x)-min(x))}
### DATASET 1 (20 observations) ###
# Response
resp1 <- lapply(data, function(x) x$Y[1:20])[[1]]
# Building the numerical variable
foo <- fda.usc::min.basis(lapply(data, function(x) x$V1)[[2]], numbasis = 15)
fd3 <- fda.usc::fdata2fd(foo$fdata.est,
type.basis = "bspline",
nbasis = foo$numbasis.opt)
foo$coef <- t(fd3$coefs)
# Building the graph variable
conn.prob <- norm01(resp1)
conn.prob[which(conn.prob < 0.05)] <- conn.prob[which(conn.prob < 0.05)] + 0.1
set.seed(12345)
graph.list <- lapply(conn.prob, function(p){sample_gnp(100, p)})
# Covariates for the full mixed model
cov.list1 <- list(lapply(data, function(x) x$V1[1:20])[[1]], graph.list, foo$coef[1:20,15])
### DATASET 2 (100 observations) ###
# Response
resp2 <- lapply(data, function(x) x$Y[1:100])[[1]]
# Building the numerical variable
foo <- fda.usc::min.basis(lapply(data, function(x) x$V1)[[2]], numbasis = 15)
fd3 <- fda.usc::fdata2fd(foo$fdata.est,
type.basis = "bspline",
nbasis = foo$numbasis.opt)
foo$coef <- t(fd3$coefs)
# Building the graph variable
conn.prob <- norm01(resp2)
conn.prob[which(conn.prob < 0.05)] <- conn.prob[which(conn.prob < 0.05)] + 0.1
set.seed(12345)
graph.list <- lapply(conn.prob, function(p){sample_gnp(100, p)})
# Covariates for the full mixed model
cov.list2 <- list(lapply(data, function(x) x$V1[1:100])[[1]], graph.list, foo$coef[1:100,15])
### DATASET 3 (200 observations) ###
# Response
resp3 <- lapply(data, function(x) x$Y)[[1]]
# Building the numerical variable
foo <- fda.usc::min.basis(lapply(data, function(x) x$V1)[[2]], numbasis = 15)
fd3 <- fda.usc::fdata2fd(foo$fdata.est,
type.basis = "bspline",
nbasis = foo$numbasis.opt)
foo$coef <- t(fd3$coefs)
# Building the graph variable
conn.prob <- norm01(resp3)
conn.prob[which(conn.prob < 0.05)] <- conn.prob[which(conn.prob < 0.05)] + 0.1
set.seed(12345)
graph.list <- lapply(conn.prob, function(p){sample_gnp(100, p)})
# Covariates for the full mixed model
cov.list3 <- list(lapply(data, function(x) x$V1)[[1]], graph.list, foo$coef[,15])
# functions ---------------------------------------------------------------
## distances & expansions inside
source("functions.R")
mb_cl_fun_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_fun_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_fun_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# fun_test1 ---------------------------------------------------------------
## distances inside, expansions outside
source("fun_test1.R")
mb_cl_ft1_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft1_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft1_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# fun_test2 ---------------------------------------------------------------
## variable selection distances & expansions outside
source("fun_test2.R")
mb_cl_ft2_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft2_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft2_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 10,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# fun_test3 ---------------------------------------------------------------
## (variable selection + clustering) distances & expansions outside
source("fun_test3.R")
mb_cl_ft3_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft3_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft3_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 10,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# fun_test4 ---------------------------------------------------------------
## (variable selection + clustering + expansions) distances & expansions outside
source("fun_test4.R")
mb_cl_ft4_1 <- microbenchmark(etree(response = resp1,
covariates = cov.list1,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft4_2 <- microbenchmark(etree(response = resp2,
covariates = cov.list2,
case.weights = NULL,
minbucket = 5,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
mb_cl_ft4_3 <- microbenchmark(etree(response = resp3,
covariates = cov.list3,
case.weights = NULL,
minbucket = 10,
alpha = 0.05,
R = 1000,
split.type = 'cluster',
coef.split.type = 'test',
nb = n.bas),
times = 10L)
# Comparison --------------------------------------------------------------
et_cl <- rbind(mb_cl_fun_1$time,
mb_cl_fun_2$time,
mb_cl_fun_3$time,
mb_cl_ft1_1$time,
mb_cl_ft1_2$time,
mb_cl_ft1_3$time,
mb_cl_ft2_1$time,
mb_cl_ft2_2$time,
mb_cl_ft2_3$time,
mb_cl_ft3_1$time,
mb_cl_ft3_2$time,
mb_cl_ft3_3$time,
mb_cl_ft4_1$time,
mb_cl_ft4_2$time,
mb_cl_ft4_3$time)
et_cl <- et_cl/1000000000
summary(et_cl)
save(mb_cl_fun_1,
mb_cl_fun_2,
mb_cl_fun_3,
mb_cl_ft1_1,
mb_cl_ft1_2,
mb_cl_ft1_3,
mb_cl_ft2_1,
mb_cl_ft2_2,
mb_cl_ft2_3,
mb_cl_ft3_1,
mb_cl_ft3_2,
mb_cl_ft3_3,
mb_cl_ft4_1,
mb_cl_ft4_2,
mb_cl_ft4_3,
file = 'TimeComparison_cluster.RData')
### Legend ###
# cl: cluster
# fun: distances & expansions inside
# ft1: distances inside, expansions outside
# ft2: variable selection distances & expansions outside
# ft3: (variable selection + clustering) distances & expansions outside
# ft4: (variable selection + clustering + expansions) distances & expansions outside
# _1: small dataset (20 observations)
# _2: medium dataset (100 observations)
# _3: large dataset (200 observations)
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.