R-Modeling(step 4)

[I]Regression

OLSregression

Description	Function
simple linear regression	lm(Y~X1,data)
polynomial regression	lm(Y~X1+...+I(X^2),data)
multiple linear regression	lm(Y~X1+X2+...+Xk)
multiple linear regression with interaction terms	lm(Y~X1+X2+X1:X2)

selection the optimal regression model

Function	Description
anova(fit1,fit2)	nested model fit
AIC(fit1,fit2)	statistical fit
stepAIC(fit,direction=)	stepwise method:"forward""backward"
regsubsets()	all-subsets regression

test

Function	Description
plot()	plot
qqplot()	quantile comparison
durbinWatsonTest()	Durbin-Waston
crPlots()	component and residual
ncvTest()	score test of non-constant error variance
spreadLevelPlot()	dispersion level
outlierTest()	Bonferroni outlier
avPlots()	added variable
inluencePlot()	regression
scatterplot()	enhanced scatter plot
scatterplotMatrix()	enhanced scatter plot matrix
vif()	variance expansion factor

abnormal

abnormal observation

1.outlier:outlierTest()
2.high leverage point:hat.plot()
3.strong influence point:Cook's D

improvement measures

1.delete observation point
2.variable transformation
3.addition and deletion of variables

Generalized Linear

generalized linear

glm()
Distribution Family | Default Connection Function

binomial	(link = "logit")
gaussian	(link = "identity")
gamma	(link = "inverse")
inverse.gaussian	(link = "1/mu^2")
poisson	(link = "log")
quasi	(link = "identity",variance = "constant")
quasibinomial	(link = "logit")
quasipoisson	(link = "log")

function with glm()
Function | Description

summary()	show details of the fitting model
coefficients()/coef()	list the parameters of the fitting model
confint()	give a confidence interval for the model paramenters
residuals()	list residual value of the fitting model
anova()	generate an analysis table of variance for two fitting model
plot()	generate a diagnostic map of the evaluation fitting model
predict()	forecast new datasets with a fitting model
deviance()	the deviation of the fitting model
df.residual()	the residual freedom of the fitting model

logistic

> data(Affairs,package="ARE")
> summary(Affairs)
> table(Affair$affairs)
> Affairs$ynaffair[Affairs$affairs > 0] <-1
> Affairs$ynaffair[Affairs$affairs == 0] <-0
> Affairs$naffair <- factor(Affairs$ynaffair),levels=c(0,1),labels=c("No","Yes"))
> table(Affairs$ynaffair)
> fit.full <-glm(ynaffair~gender + age + yearmarried +children + 
             religiousness + education + occupation + rating,family = binomial()
             data = Affairs)
> fit.reduced <-glm(ynaffair ~ age + yearmarried + religiousness +rating,
                      family = binomial(),data = Affairs)
> anova(fit.reduced,fit.full,test="Chisq")
> coef(fit.reduced)
> exp(coef(fit.reduced))
> testdata<-data.frame(rating=c(1,2,3,4,5),age=mean(Affair$age),
                                   yearsmarried=mean(Affairs$yearsmsrraied),
                                   religiousness=mean(Affairs$religiousness))
> teatdata
> testdata$prob <- predict(fit.reduced,newdata=testdata,type="response")
> testdata <- data.frame(rating = mean(Affair$rating),
                                    age=seq(17,57,10),
                                    yearsmarried=mean(Affair$yearsmarried),
                                    religiousness=mean(Affairs$religiousness))
> testdata
> testdata$prob<-predict(fit.reduced,newdata=testdata,type="response")
> testdata
> deviance(fit.reduced)/df.residual(fit.reduced)
> fit <- glm(ynaffair ~ age + yearmarried + religiousness + rating,
                 family = quasibinomial(),data=Affairs)
> pchisq(summary(fit.od)$dispersion * fit$df.residual,
             fit$df.residual,lower=F)

poisson

> data(breslow.dat,package="robust")
> name(breslow.dat)
> summary(breslow.dat[c(6,7,8,10)])
> opar<-par(no.readonly=TRUE)
> par(mfrow=c(1,2))
> attach(breakslow.dat)
> hist(sumY,breaks=20,xlab="Seizure Count",main="Distribution of Seizures")
> boxplot(sumY~Trt,xlab="Treatment",main="Group Comparisons")
> par(opar)
> fit<-glm(sumY~Base + Age +Trt,data=breslow.dat,family=poisson())
> summary(fit)
> coef(fit)
> exp(coef(fit))
> deviance(fit)/df.residual(fit)
> library(qcc)
> qcc.overdispersion.test(breslow.dat$sumY,type="poisson")
> fit.od<-glm(sumY~Base + Age + Trt,data=breslow.dat,
       family=quassipossion())
> summary(fit.od)
> fit glm(sumY~Base + Age +Trt,data=breslow.dat,offset=log(time),family=poisson)

[II]Cluster

1.cluster analysis step

1.choose the right variable
2.scale data
3.looking for anomalies
4.calculated distance
5.selection clustering algorithm
6.obtain one or more clustering methods
7.determine the number of classes
8.get the ultimate clustering solution
9.visualization of results
10.interpretation class
11.validation results

2.calculated distance

> data(nurtrient,package="flexclust")
> head(nutrient,4)
> d<-dist(nutrient)
> as.martrix(d)[1:4,1:4]

3.hierarchical clustering analysis

> data(nutrient,package="flexclust")
> row.name(nutrient) <-tolower(row.names(nutrient))
> nutrient.scaled<-scale(nutrient)

> d<-dist(nutrient.scaled)

> fit.average <-hclust(d,method="average")
> plot(fit.average,hang=-1,cex=.8,main="Average Linkage Clustering")
> library(Nbcluster)
> devAskNewPage(ask=TRUE)
> nc<-NbClust(nutrient,scaled,distance="euclidean",
                       min.nc=2,max.nc=15,method="average")
> table(nc$Best.n[1,])
> barplot(table(nc$Best,n[1,]),
              xlab="Number of Clusters",ylab="Number of Criteria",
              main="Number of Cluster Chosen by 26 Criteria")  
> clusters<-cutree(fit.average,k=5)
> table(clusters)
> aggregate(nutrient,by=list(cluster=clusters),median)
> aggregate(as.data.frame(nutrient.scaled),bu=list(cluster=clusters),median)
> plot(fit.average,hang=-1,cex=.8,
         main="Average Linkage Cluster\n5 Cluster Solution")
> rect.hclust(fit.average,k=5)

4.Clustering analysis

K-means clustering

> wssplot<-function(data,nc=15,seed=1234)(
> head(wine)
> df<-scale(wine[-1])
> wssplot(df)
> library(NcClust)
> set.seed(1234)
> devAskNewPage(ask=TRUE)
> nc<-NbClust(df,min.nc=2,max.nc=15,method="kmeans")
> table(nc$Best.n[1,)
> barplot(table(nc$Best,n[1,]),
              xlab="Number of Clusters",ylab="Number of Criteria",
              main="Number of Clusters Chosen by 26 Criteria")
> set.seed(1234)
> fit.km<-kmeans(df,3,nstart=25)
> fit.km$size
> fit.km$centers
> aggregate(wine[-1],by=list(cluster=fit.km$cluster),mean)

Division around the center point

> library(cluster)
> set.seed(1234)
> fit.pam<-pam(wine[-1],k=3,stand=TRUE)
> fit.pam$medoids
> clusplot(fit.pam,main="Bivariate Cluster Plot")
> ct.pam<-table(wine$Type,fit.pam$clustering)
> randIndex(ct.pam)

5.avoid non-existing classes

> library(fMultivar)
> set.seed(1234)
> df<-rnom2d(1000,rho=.5)
> df<-as.data.frame(df)
> plot(df,main="Bivariate Normal Distribution with rho=0.5")
> wssplot(df)
> library(NbClust)
> nc<-NbClust(df,min.nc=2,max.nc=15,method="kmean")
> dev.new()
> barplot(table(nc$Best,n[1,]),
              xlab="Number of Clusters",ylab="Number of Criteria",
              main="Number of Clusters Chosen by 26 Criteria")
> library(ggplot2)
> library(cluster)
> fit<-pam(df,k=2)
> df$clustering<-factor(fit$clustering)
> ggplot(data=df,aes(x=V1,y=V2,color=clustering,shape=clustering)) +
             geom_point() + ggtitle("Clustering of Bivariate Normal Data")
> plot(nc$All,index[,4],type="o",ylab="CCC",xlab="Number of clusters",col="blue")

[III]Classification

data preparation

> loc<-"http://archive.ics.uci.edu/ml/machine-learning-databases/"
> ds<-"breast-cancer-wisconsin/breast-cancer-wisconsin.data"
> url<-paste(loc,ds,sep="")
> breast<-read.table(url,sep=",",header=FALSE,na.string="?")
> names(breast)<-c("ID","clumpThickness","sizeUniformity",
                              "shapeUniformity","maginalAdhesion",
                              "singleEpitheliaCellSize","bareNuclei",
                              "blandChromatin","normalNucleoli","mitosis","class")
> df<-breast[-1]
> df$class<-factor(df$class,levels=c(2,4),
                           labels=c("benign","malignant"))
> set.seed(1234)
> train<-sample(nrow(df),0.7*nrow(df))
> df.train<-df[train,]
> df.validate<-df[-train,]
> table(df.train$class)
> table(df.validate$class)

logistic regression

> fit.logit<-glm(class~.,data=df.train,family=binomial())
> summary(fit.logit)
> prob<-predict(fit.logit,df.validate,type="response")
> logit.perd<-factor(prob>.5,levels=c(FALSE,TRUE),
                             labels=c("benign","malignant"))
> logit.perf<-table(df.validate$class,logit.pred,
                            dnn=c("Actual","Predicted"))
> logit.perf

decision tree

classic decision tree

> library(repart)
> set.seed(1234)
> dtree<-repart(class~.,data=df.train,method="class",
                       parms=list(split="information"))
> dtree$cptable
> plotcp(dtree)
> dtree.pruned<-prune(dtree,cp=.0125)
> library(repart.plot)
> prp(dtree.pruned,type=2,extra=104,
        fallen.leaves=TRUE,main="Decesion Tree")
> dtree.pred<-predict(dtree.pruned,df.validate,type="class")
> dtree.perf<-table(df.validate$class,dtree.pred,
                             dnn=c("Actual","Predicted"))
> dtree.perf

conditional inference tree

> library(party)
> fit.ctree<-ctree(class~.,data=sf.train)
> plot(fit.ctree,main="Conditional Inference Tree")
> ctree.pred<-predict(fit.ctree,df.validate,type="response")
> ctree.perf<-table(df.validate$class,ctree.pred
                            dnn=c("Actual","Predicted"))
> stree.perf

random forest

> library(randomForest)
> set.seed(1234)
> fit.forest<-randomForest(class~.,data=df.train,na.action=na.roughfix,importance=TRUE)
> fit.forest
> importance(fit.forest,type=2)
> forest.pred<-predict(fit.forest,df.validate)
> forest.perf<-table(df.validate$class,forest.pred,
                              dnn=c("Actual","Predicted"))
> forest.perf

support vector machines

> library(e1071)
> set.seed(1234)
> fit.svm<-svm(class~.,data=df.train)
> fit.svm
> svm,pred<-predict(fit.svm,na.omit(df.validate))
> svm.perf<-table(na.omit(df.validate)$class,
                           svm,pred,dnn=c("Actual","Predicted"))
> svm.perf

svm model with rbf core

> set.seed(1234)
> tuned<-tune.svm(class~.,data=df.train,gamma=10^(-6:1),cost=10^(-10:10))
> turned
> fit.svm<-svm(class~.,data=df.train,gamma=.01,cost=1)
> svm.pred<-predict(fit.svm,na.omit(df,validate))
> svm.perf<-table(na.omit(df.validate)$class,
                            svm.pred,dnn=c("Actual","Predicted"))
> svm.perf

choose the best solution for forecasting

> performance<-function(table,n=2){
        if(!all(dim(table)==c(2,2)))
              stop("Must be a 2 × 2 table")
   tn=table[1,1]
   fp=table[1,2]
   fn=table[2,1]
   tp=table[2,2]
   sensitivity=tp/(tp+fn)
   specificity=tn/(tn+fp)
   ppp=tp/(tp+fp)
   npp=tn/(tn+fn)
   hitrate=(tp+tn)/(tp+tn+fp+fn)
   result<-paste("Sensitivity=",round(ppp,n),
           "\nSpecificity=",round(specificity,n),
           "\nPositive Predictive Value=",round(ppp,n),
           "\nNegative Predictive Value=",round(npp,n),
           "\nAccuracy=",round(hitrate,n),"\n",seq="")
   cat(result)
  }

data mining with the rattle package

> install.packages("rattle")
> rattle()
> loc<-"http://archive.ics.uci.edu/ml/machine-learning-databases/"
> ds<-"pima-indians-diabetes/pima-indians-diabetes.data"
> url <-paste(loc,ds,sep="")
> diabetes<-read.table(url,seq=",",header=FALSE)
> names(diabetes)<-c("npregant","plasma","bp","triceps",
                                  "insulin","bmi","pedigree","age","class")
> diabetes$class<-factor(diabetes$class,levels=c(0,1),
                                      labels=c("normal","diabetic"))
> library(rattle)
> rattle()
> cv<-matrix(c(145,50,8,27),nrow=2)
> performance(as.table(cv))

[IV]Time Series

Function	Package	Description
ts()	stats	generate timing objects
plot()	graphics	draw a line graph of the time series
start()	stats	return the start time of the time series
end()	stats	return the end time of the time series
frequency()	stats	return the number of time points of the time series
windows()	stats	subset a sequence object
ma()	forecast	fit a simple moving average model
stl()	stats	use LOESS smooth to decompose timing into seasonal terms,trend terms and random terms
monthplot()	stats	draw seasonal terms in time series
seasonplot() forecast	generate seasonal plot
HoltWinters()	ststs	fit exponential smoothing model
forecast()	forecast	predicting the future value of the timing
accuracy()	forecast	return time-of-fit goodness variable
ets()	forecast	fit exponential smoothing model and automatically select the optimal model
lag()	stats	return the timing after the specified hysteresis is taken
Acf()	forecast	estimated autocorrelation function
Pacf()	forecast	estimated partial autocorrelation function
diff()	base	return the sequence after the lag term or difference
ndiffs()	forecast	find the optimal difference count to remove the trend terms in the sequence
adf.test()	tseries	perform an ADF test on the sequence to determine if it is stable
arima	stats	fit the ARIMA model
Box.test()	ststs	perform a Ljung-Box test to determine if the model's residuals are independent
bds.test()	teeries	perform a BDS test to determine whether random variables in the sequence are subject to independent and identical distribution
auto.arima	forecast	automatically select the ARIMA model

END!