xwMOOC 기계학습
사례 - 콘크리트 강도 데이터
학습목표
- UCI 콘크리트 강도 데이터로 예측모형을 개발한다.
1. UCI 콘크리트 데이터
UCI 콘크리트 강도 데이터는 Concrete Compressive Strength Data Set 데이터를 사용해서 강도를 예측한다.
- Name – Data Type – Measurement – Description
- Cement (component 1) – quantitative – kg in a m3 mixture – Input Variable
- Blast Furnace Slag (component 2) – quantitative – kg in a m3 mixture – Input Variable
- Fly Ash (component 3) – quantitative – kg in a m3 mixture – Input Variable
- Water (component 4) – quantitative – kg in a m3 mixture – Input Variable
- Superplasticizer (component 5) – quantitative – kg in a m3 mixture – Input Variable
- Coarse Aggregate (component 6) – quantitative – kg in a m3 mixture – Input Variable
- Fine Aggregate (component 7) – quantitative – kg in a m3 mixture – Input Variable
- Age – quantitative – Day (1~365) – Input Variable
- Concrete compressive strength – quantitative – MPa – Output Variable
예측변수 및 종속변수 모두 연속형 변수이며 결측값도 없고 전형적인 공학 데이터 형태를 띄고 있다.
2. 콘크리트 데이터 예측모형 적용
데이터는 AppliedPredictiveModeling 팩키지에 concrete로 저장되어 있고, Applied Predictive Modeling 책에 적용된 모형을 그대로 실행한다.
mixtures 데이터셋은 concrete 데이터의 짝꿍으로 이해하면 된다.
# 0.0 환경설정, 각자 컴퓨터 코어 숫자를 확인하고 설정한다.
# 기본원칙은 최대 코어숫자에서 1개를 뺀 숫자를 작업에 할당한다.
suppressMessages(library(doMC))
registerDoMC(cores=7)
# 1.1. 데이터 불러오기
suppressMessages(library(caret))
suppressMessages(library(AppliedPredictiveModeling))
suppressMessages(library(dplyr))
suppressMessages(library(plyr))
data(concrete)
str(concrete)
# 1.2. 훈련데이터, 검증데이터
averaged <- ddply(mixtures,
.(Cement, BlastFurnaceSlag, FlyAsh, Water,
Superplasticizer, CoarseAggregate,
FineAggregate, Age),
function(x) c(CompressiveStrength =
mean(x$CompressiveStrength)))
forTraining <- createDataPartition(averaged$CompressiveStrength,
p = 3/4)[[1]]
trainingSet <- averaged[ forTraining,]
testSet <- averaged[-forTraining,]
##==============================================================================================
## 02. EDA
##==============================================================================================
# describe in Hmisc
# featurePlot in caret
featurePlot(x = concrete[, -9],
y = concrete$CompressiveStrength,
between = list(x = 1, y = 1),
type = c("g", "p", "smooth"))
##==============================================================================================
## 03. 예측 모형 적합
##==============================================================================================
# 모형 공식 생성
modFormula <- paste("CompressiveStrength ~ (.)^2 + I(Cement^2) + ",
"I(BlastFurnaceSlag^2) + I(FlyAsh^2) + I(Water^2) +",
" I(Superplasticizer^2) + I(CoarseAggregate^2) + ",
"I(FineAggregate^2) + I(Age^2)")
modFormula <- as.formula(modFormula)
##==============================================================================================
## 04. 컴퓨팅 기반 예측모형
##==============================================================================================
#------------------------------------------------------------------------------------
# 제어변수 설정
controlObject <- trainControl(method = "repeatedcv",
repeats = 1, number = 1,
verboseIter = TRUE)
# 1. 선형회귀
linearReg <- train(modFormula,
data = trainingSet,
method = "lm",
trControl = controlObject)
linearReg
# 2. 부분최소자승
library(pls)
plsModel <- train(modFormula, data = trainingSet,
method = "pls",
preProc = c("center", "scale"),
tuneLength = 15,
trControl = controlObject)
plsModel
# 3. 부분 최소자승회귀
library(elasticnet)
enetGrid <- expand.grid(.lambda = c(0, .001, .01, .1),
.fraction = seq(0.05, 1, length = 20))
enetModel <- train(modFormula, data = trainingSet,
method = "enet",
preProc = c("center", "scale"),
tuneGrid = enetGrid,
trControl = controlObject)
enetModel
# 4. MARS
earthModel <- train(CompressiveStrength ~ ., data = trainingSet,
method = "earth",
tuneGrid = expand.grid(.degree = 1,
.nprune = 2:25),
trControl = controlObject)
earthModel
# 5. SVM
svmRModel <- train(CompressiveStrength ~ ., data = trainingSet,
method = "svmRadial",
tuneLength = 15,
preProc = c("center", "scale"),
trControl = controlObject)
svmRModel
# 6. 신경망
nnetGrid <- expand.grid(.decay = c(0.001, .01, .1),
.size = seq(1, 27, by = 2),
.bag = FALSE)
nnetModel <- train(CompressiveStrength ~ .,
data = trainingSet,
method = "avNNet",
tuneGrid = nnetGrid,
preProc = c("center", "scale"),
linout = TRUE,
trace = FALSE,
maxit = 1000,
trControl = controlObject)
nnetModel
# 7. 나무 모형(Tree Model)
rpartModel <- train(CompressiveStrength ~ .,
data = trainingSet,
method = "rpart",
tuneLength = 30,
trControl = controlObject)
ctreeModel <- train(CompressiveStrength ~ .,
data = trainingSet,
method = "ctree",
tuneLength = 10,
trControl = controlObject)
mtModel <- train(CompressiveStrength ~ .,
data = trainingSet,
method = "M5",
trControl = controlObject)
# 8. 나무 투표 모형
treebagModel <- train(CompressiveStrength ~ .,
data = trainingSet,
method = "treebag",
trControl = controlObject)
treebagModel
rfModel <- train(CompressiveStrength ~ .,
data = trainingSet,
method = "rf",
tuneLength = 10,
ntrees = 1000,
n.minobsinnode = 10,
importance = TRUE,
trControl = controlObject)
rfModel
gbmGrid <- expand.grid(.interaction.depth = seq(1, 7, by = 2),
.n.trees = seq(100, 1000, by = 50),
.n.minobsinnode = 10,
.shrinkage = c(0.01, 0.1))
gbmModel <- train(CompressiveStrength ~ .,
data = trainingSet,
method = "gbm",
tuneGrid = gbmGrid,
verbose = TRUE,
trControl = controlObject)
cubistGrid <- expand.grid(.committees = c(1, 5, 10, 50, 75, 100),
.neighbors = c(0, 1, 3, 5, 7, 9))
cbModel <- train(CompressiveStrength ~ .,
data = trainingSet,
method = "cubist",
tuneGrid = cubistGrid,
verbose = TRUE,
trControl = controlObject)
#------------------------------------------------------------------------------------
# 모형평가 시각화
allResamples <- resamples(list("Linear Reg" = linearReg,
"PLS" = plsModel,
"Elastic Net" = enetModel,
MARS = earthModel,
SVM = svmRModel,
"Neural Networks" = nnetModel,
CART = rpartModel,
"Cond Inf Tree" = ctreeModel,
"Bagged Tree" = treebagModel,
"Boosted Tree" = gbmModel,
"Random Forest" = rfModel,
Cubist = cbModel))
bwplot(allResamples, layout = c(2, 1),scales = list(relation = "free"),
xlim = list(c(0, 10), c(0.5, 1)))