Arvores de decisao condicional
library(tidyverse)
## ── Attaching packages ────────────────────────────────── tidyverse 1.3.2.9000 ──
## ✔ ggplot2 3.3.6 ✔ dplyr 1.0.10
## ✔ tibble 3.1.8 ✔ stringr 1.4.1
## ✔ tidyr 1.2.1 ✔ forcats 0.5.2
## ✔ readr 2.1.3 ✔ lubridate 1.8.0
## ✔ purrr 0.3.5
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
library(readr)
setwd("/Users/daniloap/Library/CloudStorage/OneDrive-Pessoal/Laboratorio_BKP/Pedro Brandao/PDCRS/")
roc <- read_delim("pdcrsROC.csv",
delim = ";", escape_double = FALSE, trim_ws = TRUE)
## Rows: 69 Columns: 16
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ";"
## dbl (16): grupo.clinico, PDCRS, Idade, Escolaridade, FrontalSub, PosteriorCo...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
library(caret)
## Loading required package: lattice
##
## Attaching package: 'caret'
##
## The following object is masked from 'package:purrr':
##
## lift
# Define o controle repeated 10 k-fold
train_control <- trainControl(method = "repeatedcv", number = 10, repeats = 3)
roc$group <- ifelse(roc$grupo.clinico == 0, "NC", "MCI")
roc$group <- as.factor(roc$group)
# train the model
model <- train(group ~ PDCRS + MoCA,
data = roc, trControl = train_control, method = "ctree")
# summarize results
print(model)
## Conditional Inference Tree
##
## 69 samples
## 2 predictor
## 2 classes: 'MCI', 'NC'
##
## No pre-processing
## Resampling: Cross-Validated (10 fold, repeated 3 times)
## Summary of sample sizes: 62, 62, 63, 61, 63, 62, ...
## Resampling results across tuning parameters:
##
## mincriterion Accuracy Kappa
## 0.01 0.7220238 0.3420859
## 0.50 0.7303571 0.3563716
## 0.99 0.7351190 0.3572499
##
## Accuracy was used to select the optimal model using the largest value.
## The final value used for the model was mincriterion = 0.99.
plot(model)
model0 <- train(group ~ ImmediateMemo + Naming + SustAttention + WorkingMemo +
Clock + ClockCopy + DelayedMemo + AlternatedFluency +
ActionFluency, data = roc,
trControl = train_control, method = "ctree")
# summarize results
print(model0)
## Conditional Inference Tree
##
## 69 samples
## 9 predictor
## 2 classes: 'MCI', 'NC'
##
## No pre-processing
## Resampling: Cross-Validated (10 fold, repeated 3 times)
## Summary of sample sizes: 62, 63, 62, 62, 63, 61, ...
## Resampling results across tuning parameters:
##
## mincriterion Accuracy Kappa
## 0.01 0.6984127 0.3189713
## 0.50 0.6888889 0.3064072
## 0.99 0.6301587 0.1777797
##
## Accuracy was used to select the optimal model using the largest value.
## The final value used for the model was mincriterion = 0.01.
plot(model0)
library(partykit)
## Loading required package: grid
## Loading required package: libcoin
## Loading required package: mvtnorm
ct <- ctree(group ~ ImmediateMemo + Naming + SustAttention + WorkingMemo +
Clock + ClockCopy + DelayedMemo + AlternatedFluency +
ActionFluency, data = roc, mincriterion = 0.95)
ct
##
## Model formula:
## group ~ ImmediateMemo + Naming + SustAttention + WorkingMemo +
## Clock + ClockCopy + DelayedMemo + AlternatedFluency + ActionFluency
##
## Fitted party:
## [1] root
## | [2] ImmediateMemo <= 9
## | | [3] AlternatedFluency <= 12: MCI (n = 26, err = 26.9%)
## | | [4] AlternatedFluency > 12: NC (n = 7, err = 0.0%)
## | [5] ImmediateMemo > 9: NC (n = 36, err = 16.7%)
##
## Number of inner nodes: 2
## Number of terminal nodes: 3
plot(ct)
#png("ctree.01.png", res=300, height=8, width=14, units="in")
#plot(as.simpleparty(ct))
#dev.off()
##
## Model formula:
## group ~ PDCRS
##
## Fitted party:
## [1] root
## | [2] PDCRS <= 77: MCI (n = 11, err = 0.0%)
## | [3] PDCRS > 77
## | | [4] PDCRS <= 96: NC (n = 32, err = 40.6%)
## | | [5] PDCRS > 96: NC (n = 26, err = 3.8%)
##
## Number of inner nodes: 2
## Number of terminal nodes: 3
##
## Model formula:
## group ~ PDCRS + MoCA
##
## Fitted party:
## [1] root
## | [2] MoCA <= 21: MCI (n = 14, err = 7.1%)
## | [3] MoCA > 21
## | | [4] PDCRS <= 95: NC (n = 26, err = 42.3%)
## | | [5] PDCRS > 95: NC (n = 29, err = 3.4%)
##
## Number of inner nodes: 2
## Number of terminal nodes: 3
##
## Model formula:
## group ~ ImmediateMemo + Naming + SustAttention + WorkingMemo +
## Clock + ClockCopy + DelayedMemo + AlternatedFluency + ActionFluency
##
## Fitted party:
## [1] root
## | [2] ClockCopy <= 9
## | | [3] SustAttention <= 8: MCI (n = 8, err = 0.0%)
## | | [4] SustAttention > 8: MCI (n = 12, err = 50.0%)
## | [5] ClockCopy > 9
## | | [6] WorkingMemo <= 6: NC (n = 17, err = 47.1%)
## | | [7] WorkingMemo > 6
## | | | [8] ActionFluency <= 17: NC (n = 12, err = 25.0%)
## | | | [9] ActionFluency > 17: NC (n = 20, err = 0.0%)
##
## Number of inner nodes: 4
## Number of terminal nodes: 5
##
## Model formula:
## group ~ ImmediateMemo + Naming + SustAttention + WorkingMemo +
## Clock + DelayedMemo + AlternatedFluency + ActionFluency
##
## Fitted party:
## [1] root
## | [2] DelayedMemo <= 5
## | | [3] WorkingMemo <= 6: MCI (n = 18, err = 11.1%)
## | | [4] WorkingMemo > 6: NC (n = 14, err = 28.6%)
## | [5] DelayedMemo > 5
## | | [6] ActionFluency <= 17: NC (n = 17, err = 29.4%)
## | | [7] ActionFluency > 17: NC (n = 20, err = 0.0%)
##
## Number of inner nodes: 3
## Number of terminal nodes: 4
##
## Model formula:
## group ~ MoCA + ImmediateMemo + Naming + SustAttention + WorkingMemo +
## Clock + ClockCopy + DelayedMemo + AlternatedFluency + ActionFluency
##
## Fitted party:
## [1] root
## | [2] MoCA <= 21: MCI (n = 14, err = 7.1%)
## | [3] MoCA > 21
## | | [4] WorkingMemo <= 6: NC (n = 18, err = 44.4%)
## | | [5] WorkingMemo > 6
## | | | [6] Naming <= 17: NC (n = 12, err = 25.0%)
## | | | [7] Naming > 17
## | | | | [8] MoCA <= 27: NC (n = 18, err = 0.0%)
## | | | | [9] MoCA > 27: NC (n = 7, err = 14.3%)
##
## Number of inner nodes: 4
## Number of terminal nodes: 5
##
## Model formula:
## group ~ FrontalSub + PosteriorCort
##
## Fitted party:
## [1] root
## | [2] FrontalSub <= 58: MCI (n = 22, err = 22.7%)
## | [3] FrontalSub > 58
## | | [4] FrontalSub <= 74: NC (n = 26, err = 30.8%)
## | | [5] FrontalSub > 74: NC (n = 21, err = 0.0%)
##
## Number of inner nodes: 2
## Number of terminal nodes: 3
