# seminar 2 ---------------------------------------------------------------
library(corrplot)
library(ggplot2)
library(Hmisc)
library(PerformanceAnalytics)
library( scatterplot3d)
library(FactoMineR)
library(factoextra)
library(cluster)
library(ca)
datepath<-"D:\\cursuri\\master\\data mining\\proiect"
date<-read.csv("indicatori.csv")
date<-read.table(file.path(datepath,"indicatori.csv"),sep=",",header=TRUE, row.names=1)
View(date)
summary(date)
a<-sd(date$I1)
b<-sd(date$I2)
c<-sd(date$I3)
d<-sd(date$I4)
e<-sd(date$I5)
f<-sd(date$I6)
g<-sd(date$I7)
h<-sd(date$I8)
s=c(a,b,c,d,e,f,g,h)
View(s)
boxplot(date,col=c("red","green","blue","yellow","ivory","pink","orange","purple"))
corrplot( c2$r, type="upper", p.mat=c2$P, sig.level= 0.05, insig="blank")
chart.Correlation(date,histogram=TRUE, pch=19)
indic<- read.csv("indicatori.csv")
ind<-indic[2:9]
View(ind)
acp1<- princomp( ind, cor= TRUE, scores= TRUE)
summary(acp1)
plot(acp1, type="l")
biplot(acp1)
acp1$loadings
scoruri<- acp1$ scores
scoruri
acp2<-prcomp(ind, center= TRUE, scale= TRUE)
summary(acp2)
plot(acp2, type="l")
biplot(acp2)
scoruri2<- acp2$ scores
cor(ind, scoruri2[,1:2])
acp3<- PCA(ind)
summary(acp3)
fviz_pca_var(acp3,axes=c(1,2),col.var="contrib")
plot(acp3,choix="ind")
varimax<- varimax(acp2$rotation[,1:2])
scorurirotite<- scale(ind)%*% varimax$loadings
cor(ind,scorurirotite)
plot(scorurirotite[,1],scorurirotite[,2])
d<- dist(scorurirotite[,1:2],method = "euclidean")
sol1<- hclust(d, method = "ward.D2")
plot(sol1)
sol1.2<-cutree(sol1, k=2)
s1.2<-silhouette(sol1.2, d )
plot(s1.2)
sol2<-kmeans(scorurirotite[,1:2],2) # 2 de la final ne spune cate clustere alegem
s2.2<-silhouette(sol2$cluster,d)
plot(s2.2)
fviz_cluster(list(data=scorurirotite[,1:2],cluster=sol2$cluster))# pt Kmeans
fviz_cluster(list(data=scorurirotite[,1:2],cluster=cutree(sol1, k=2))) # ceva nu e in regula
rasp<- read.csv("formular3.csv")
View(rasp)
attach(rasp)
rasp2<- subset(rasp,Solutie2!="")
View(rasp2)
rasp2$Solutie1<- rasp2$Solutie2
rasp3<- rbind(rasp,rasp2)
View(rasp3)
fix(rasp3)
tabel<-table(rasp3$Solutie1, rasp3$durata)
View(tabel)
levels(rasp3$Solutie1)
levels(rasp3$Solutie1)[levels (rasp3$Solutie1)%in%c("Tonomat")]<-"Cumpar de la supermarket"
levels(rasp3$Solutie1)
tabel<-table(rasp3$Solutie1, rasp3$durata)
View(tabel)
tabel<-tabel[-5,]
View(tabel)
corespond<- ca(tabel)
summary(corespond)
window()
plot(corespond)
library(ggplot2)
library(memisc)
library(ROCR)
library(ROSE)
# regresie logistica ------------------------------------------------------
a<- data.frame(as.data.set(spss.system.file("antrep.sav")))
View(a)
attach(a)
table(country)
China<-subset(a,country=="China")
table(China$bstart)
variable<-which(names(China)%in%c("bstart","gemwork3","gemhhinc","gemeduc","knowent","suskill","fearfail","gender","age9c"))
variable
China<-China[,variable]
China<-na.omit(China)
China
table(China$bstart)
model1<-glm(bstart~gemwork3+gemhhinc+knowent+gender,data=China,family = "binomial")
model1
summary(model1)
exp(coef(model1))
table(China$gemeduc)
levels(China$gemeduc)[levels(China$gemeduc) %in% c("NONE","SOME SECONDARY")] <- "level1"
levels(China$gemeduc)[levels(China$gemeduc) %in% c("SECONDARY DEGREE")] <- "level2"
levels(China$gemeduc)[levels(China$gemeduc) %in% c("POST SECONDARY","GRAD EXP")] <- "level3"
table(China$gemeduc)
table(China$age9c)
levels(China$age9c)[levels(China$age9c) %in% c("0-17","18-24")]<- "categorie1"
levels(China$age9c)[levels(China$age9c) %in% c("25-34","35-44","45-54")]<- "categorie2"
levels(China$age9c)[levels(China$age9c) %in% c("55-64","65-120")]<- "categorie3"
table(China$age9c)
model3<-glm(bstart~gemwork3+knowent+suskill, data=China, family="binomial")
summary(model3)
exp(coef(model3))
library(ROCR)
yhat<- predict(model3, type="response")
pr<-prediction(yhat, China$bstart, label.ordering = NULL)
perf<-performance(pr, "tpr", "fpr")
windows()
plot(perf, colorize= TRUE, lwd=5)
performance(pr, "auc")
library(ROSE)
table(China$bstart)
date.rose<-ROSE(bstart~.,data=China,p=0.4, seed = 123) #p arata cum reechilibrez
table(date.rose$data$bstart)
library(Matching)
date.rose<-ROSE(bstart~.,data=China,p=0.4, seed = 123)$data #p arata cum reechilibrez
table(date.rose$bstart)
model<-glm(bstart~ gemwork3+knowent+suskill,data=date.rose,family = "binomial")
summary(model)
exp(coef(model))
library(ROCR)
yhat<- predict(model, type="response")
pr<-prediction(yhat, date.rose$bstart, label.ordering = NULL)
perf<-performance(pr, "tpr", "fpr")
plot(perf, colorize= TRUE, lwd=5)
performance(pr, "auc")
country<- bstart~gemwork3+knowent+suskill
country1<-country[,c(1,2,4,7,8)] # aici am pastrat clasificatorii de mai sus, sunt cei folositi in model2
View(country1)
View(country)
perfmRose<- ROSE.eval(bstart~., data= country1,
                      learner = glm, method.assess = "LKOCV", K=5,
                      control.learner = list(family=binomial),
                      control.rose=list(p=0.4),seed=123, trace=TRUE)
summary(perfmRose)
data("lalonde")
View(lalonde)
dim(lalonde)
table(lalonde$treat)
head(lalonde)
mean(lalonde$re78[lalonde$treat== 1])
mean(lalonde$re78[lalonde$treat== 0])
scor<- glm(treat~ age + educ+ black+ hisp+ married+ nodegr, data= lalonde, family=binomial)
summary(scor)
efect<-Match(Y=lalonde$re78, Tr= lalonde$treat, X= scor$fitted, estimand = "ATT", M=1, replace= TRUE)
summary(efect)
MatchBalance(treat~ age + educ+ black+ hisp+ married+ nodegr, match.out= efect, data=lalonde, nboots = 200)
qqplot(lalonde$age[efect$index.control], lalonde$age[efect$index.treated])
abline(coef=c(0,1), col=2)
scor<- glm(treat~ educ+ black+ hisp+ married+ nodegr, data= lalonde, family=binomial)
efect<-Match(Y=lalonde$re78, Tr= lalonde$treat, X= scor$fitted, estimand = "ATT", M=1, replace= TRUE)
summary(efect)
# ne uitam cat de mult difera fata de primul
MatchBalance(treat~ educ+ black+ hisp+ married+ nodegr, match.out= efect, data=lalonde, nboots = 200)
qqplot(lalonde$educ[efect$index.control], lalonde$educ[efect$index.treated])
abline(coef=c(0,1), col=2)
#adaugam age si eliminam black
scor<- glm(treat~ age+educ+ hisp+ married+ nodegr, data= lalonde, family=binomial)
efect<-Match(Y=lalonde$re78, Tr= lalonde$treat, X= scor$fitted, estimand = "ATT", M=1, replace= TRUE)
summary(efect)
MatchBalance(treat~ age+educ+ hisp+ married+ nodegr, match.out= efect, data=lalonde, nboots = 200)
qqplot(lalonde$hisp[efect$index.control], lalonde$hisp[efect$index.treated])
abline(coef=c(0,1), col=2)
#adaugam black si eliminam bhisp
scor<- glm(treat~ age+educ+ black+ married+ nodegr, data= lalonde, family=binomial)
efect<-Match(Y=lalonde$re78, Tr= lalonde$treat, X= scor$fitted, estimand = "ATT", M=1, replace= TRUE)
summary(efect)
MatchBalance(treat~ age+educ+ black+ married+ nodegr, match.out= efect, data=lalonde, nboots = 200)
qqplot(lalonde$black[efect$index.control], lalonde$black[efect$index.treated])
abline(coef=c(0,1), col=2)
#adaugam black si eliminam hisp & age
scor<- glm(treat~ educ+ black+ married+ nodegr, data= lalonde, family=binomial)
efect<-Match(Y=lalonde$re78, Tr= lalonde$treat, X= scor$fitted, estimand = "ATT", M=1, replace= TRUE)
summary(efect)
MatchBalance(treat~ educ+ black+ married+ nodegr, match.out= efect, data=lalonde, nboots = 200)
qqplot(lalonde$black[efect$index.control], lalonde$black[efect$index.treated])
abline(coef=c(0,1), col=2)
#adaugam black si eliminam hisp & age & educ
scor<- glm(treat~ black+ married+ nodegr, data= lalonde, family=binomial)
efect<-Match(Y=lalonde$re78, Tr= lalonde$treat, X= scor$fitted, estimand = "ATT", M=1, replace= TRUE)
summary(efect)
MatchBalance(treat~  black+ married+ nodegr, match.out= efect, data=lalonde, nboots = 200)
qqplot(lalonde$black[efect$index.control], lalonde$black[efect$index.treated])
abline(coef=c(0,1), col=2)
#adaugam black si hisp si eliminam & age & educ
scor<- glm(treat~ hisp+black+ married+ nodegr, data= lalonde, family=binomial)
efect<-Match(Y=lalonde$re78, Tr= lalonde$treat, X= scor$fitted, estimand = "ATT", M=1, replace= TRUE)
summary(efect)
MatchBalance(treat~ hisp+black+ married+ nodegr, match.out= efect, data=lalonde, nboots = 200)
qqplot(lalonde$black[efect$index.control], lalonde$black[efect$index.treated])
abline(coef=c(0,1), col=2)
#sa mai testam si chestia urmatoare, alta metoda gen
library(rgenoud)
x<-cbind(lalonde$hisp, lalonde$black, lalonde$married, lalonde$nodegr)
genetic<- GenMatch(Tr=lalonde$treat, X=x, pop.size = 1000)
#daca crestem sau scadem pop.size se modifica, cu cat e mai mare cu atat solutia e mai buna
mgen1<-Match(Y=lalonde$re78, Tr=lalonde$treat, X=x, Weight.matrix = genetic)
summary(mgen1)
MatchBalance(treat~ hisp+black+ married+ nodegr, match.out= mgen1, data=lalonde, nboots = 200)
# Analiza conjoint --------------------------------------------------------
library(support.CEs)
library(AlgDesign)
ffd<-gen.factorial(c(2,2,2,2), varNames = c("Memorie","Rezolutie","Diagonala","Pret"),factors = "all")
View(ffd)
#eliminam 8(1112)si 9(2221)
ffd<-ffd[-c(1,4,6,8,9,11,15,16),]r
set.seed(100)
des<-rotation.design(candidate.array = ffd, attribute.names = list(Memorie=c("sub 32","peste 32"), Rezolutie=c("sub 12","peste 12"),
                                                                   Diagonala=c("sub 5.5", "peste 5.5"), Pret=c("sub 2000","peste 2000")),
                     nalternatives = 2, nblocks=1, row.renames= FALSE, randomize= FALSE, seed=100)
des
desmat<-make.design.matrix(choice.experiment.design = des, optout = FALSE,
                           categorical.attributes = c("Memorie","Rezolutie","Diagonala","Pret"),
                           unlabeled = TRUE)
q<-questionnaire(choice.experiment.design = des)
rasp<-read.table("Raspunsuri.txt",header=TRUE, sep= "\t")
View(rasp)
dataset<- make.dataset(respondent.dataset = rasp, choice.indicators = c("I1","I2","I3","I4","I5","I6","I7","I8"),
                       design.matrix = desmat)
library(survival)
rezultat<-clogit(RES~ASC+peste.32+peste.12+peste.5.5+peste.2000 + strata(STR),data=dataset)
summary(rezultat)
library(rpart)
library(rpart.plot)
library(ROSE)
library(ROCR)
library(memisc)
a<- data.frame(as.data.set(spss.system.file("antrep.sav")))
attach(a)
China<-subset(a,country=="China")
variable<-which(names(China)%in%c("bstart","gemwork3","gemhhinc","gemeduc","knowent","suskill","fearfail","gender","age9c"))
variable
China<-China[,variable]
China<-na.omit(China)
China
date.rose<-ROSE(bstart~.,data=China,p=0.4, seed = 123)$data #p arata cum reechilibrez
model<-rpart(bstart~gemwork3+knowent+suskill, method="class", data=date.rose)
summary(model)
datarose<-ROSE(bstart~gemwork3+knowent+suskill , data=date.rose , p=0.4, seed=123)$data
prp(model, type=3, extra=106, under=TRUE, box.palette = "BuPu")
plotcp(model)
printcp(model)
arboref<-prune(model, cp=0.04)
arboref
prp(arboref)
esantion<- sample(1:nrow(date.rose), 1130)
train<-date.rose[esantion,]
test<-date.rose[-esantion,]

p<- predict(model, test, type="prob")[,2] #prob - probabilitatea pe clasa 1 si pe clasa 2, [,2]- pentru a alege a 2a clasa, YES
pred<-prediction(p, test$bstart)
roc<-performance(pred, "tpr", "fpr")
plot(roc, colorize = TRUE, lwd=5)
performance(pred, "auc")  #auc=0.6749
eval<-ROSE.eval(bstart~., data=country, learner = rpart,
                method.assess = "LKOCV", K=5, extr.pred = function(obj)obj[,2],
                control.rose = list(p=0.4),
                control.learner=list(control=list(cp=0.04)), seed=123, trace=TRUE)
modell<-glm(bstart~ gemwork3+knowent+suskill,data=date.rose,family = "binomial")
summary(modell)
exp(coef(modell))
library(ROCR)
yhat<- predict(modell, type="response")
pr<-prediction(yhat, date.rose$bstart, label.ordering = NULL)
perf<-performance(pr, "tpr", "fpr")
plot(perf, colorize= TRUE, lwd=5)
performance(pr, "auc")

#arbore simplu
model<-rpart(bstart~gemwork3+knowent+suskill, method="class", data=date.rose)
p<- predict(model, test, type="prob")[,2] #prob - probabilitatea pe clasa 1 si pe clasa 2, [,2]- pentru a alege a 2a clasa, YES
pred<-prediction(p, test$bstart)
roc<-performance(pred, "tpr", "fpr")
plot(roc, colorize = TRUE, lwd=5)
performance(pred, "auc")


predl<-predict(modell,test, type="response")
predarb<-predict(model,test, type="prob")[,2]
pred<-cbind(predl, predarb)
prednot<-prediction(pred, cbind(test$bstart, test$bstart))
perf<-performance(prednot,"tpr","fpr")
plot(perf, col=as.list(1:2))
legend("bottomright", legend=c("logistica","arbore"),pch=19,col=c("black","red"))


# random forest  ----------------------------------------------------------

library(randomForest)
View(train[,7])
rf<-randomForest(x=train[,-7],y=train$bstart, importance= TRUE, ntree=1000, replace= TRUE,
                 xtest=test[,-7], ytest=test$bstart)

rf$importance
library(ROCR)
p<- predict(rf, test, type="prob")#prob - probabilitatea pe clasa 1 si pe clasa 2, [,2]- pentru a alege a 2a clasa, YES
pred<-prediction(p, test$bstart)
roc<-performance(pred, "tpr", "fpr")
plot(roc, colorize = TRUE, lwd=5)
performance(pred, "auc")
summary(rf)