# R programma

data(cars)
cor.test(cars$speed,cars$dist)

summary(lm(dist~speed,data=cars))

plot(cars$dist~cars$speed)
abline(lm(dist~speed,data=cars))

help.start()

help(plot)

args(cor)

example(plot)

instal.packages("ggplot2")
library(ggplot2)

4+7
log(8,2)
exp(2)

getwd()
setwd("/Users/didzis/Documents")
getwd()

dati <- read.table(file="niedres.txt",header=T, sep="\t", dec=".")
dati2 <- read.csv2(file="niedres.txt",header=T, sep="\t", dec=".")

dati

write.table(x=dati, file="eksports.txt",sep="\t",dec=".")

str(dati)


otrais <- c("A","B","C","D","E","F")
otrais

otrais.fakt<-factor(otrais)
otrais.fakt

otrais.fakt <- factor(otrais,levels=c("D","A","C","B","E","F"))
otrais.fakt


dim(dati) #datu tabula
names(dati)

head(dati)
tail(dati,n=3)

dati$garums
dati["garums"]

dati[,1]
dati[,1,drop=FALSE]

dati[1,]

dati[1,2]

dati$garums[1]
dati$garums[c(1,4)]
dati$garums[-3]

garums

ls()

rm(otrais)




#Vienas paraugkopas analize
## Paraugkopas grafiska analize

## Dati

niedr<-read.table(file="niedres2.txt", header=TRUE,sep="\t",dec=".")
str(niedr)

## Grafiska analize
plot(niedr$garums)

plot(niedr$garums,niedr$platums)

dotchart(niedr$garums)

dotchart(niedr$garums,groups=niedr$paraug,pch=as.numeric(niedr$paraug))

boxplot(niedr$garums)

boxplot(niedr$garums~niedr$paraug)

stripchart(niedr$garums,method="stack",vert=TRUE)

hist(niedr$garums)

## Atbilstiba normalajam sadalijumam
## Grafiska parbaude - histogramma
par(mfrow=c(1,2))
hist(niedr$garums,breaks=20)
hist(niedr$garums,breaks=10)

## Grafisk? p?rbaude - QQplot 1
qqnorm(niedr$garums)
qqline(niedr$garums)

## Grafiska parbaude - QQplot 2
library(car)
qqPlot(niedr$garums)


## Analitiska parbaude - Sapiro tests 1
shapiro.test(niedr$garums)




# Statistiskie raditaji 

niedr<-read.table(file="niedres2.txt", header=TRUE,sep="\t",dec=".")
str(niedr)


mean(niedr$garums)
sd(niedr$garums)
var(niedr$garums)
median(niedr$garums)

round(mean(niedr$garums),1)

x <- c(1:20,NA)
mean(x)
mean(x,na.rm=TRUE)

min(niedr$garums)
max(niedr$garums)
range(niedr$garums)

quantile(niedr$garums)
quantile(niedr$garums,probs=c(0.025,0.975))

summary(niedr$garums)

summary(niedr)




##Paraugkopu salidzinsaana
## Dati 1
niedr2<-read.table(file="lapas.txt", 
                  header=TRUE,sep="\t",dec=".")
str(niedr2)

summary(niedr2)

## Dispersiju salidzinasana 2

var.test(niedr2$garums~niedr2$paraug)

## Videjo aritmrtisko salidzinasana 2

t.test(niedr2$garums~niedr2$paraug,var.equal=TRUE)

## Saistitu paraugkopu videjo aritmetisko salidzinasana 2

rokas <- read.table(file="rokas.txt",header=TRUE,
                    sep="\t",dec=".")
str(rokas)

t.test(rokas$laba,rokas$kreisa,paired=TRUE)


## T-tests vienai paraugkopai 2

t.test(niedr2$platums,mu=3.0)

## Neparametriskas metodes - neatkarigas paraugkopas 1

wilcox.test(niedr2$garums~niedr2$paraug)

## Neparametriskas metodes - atkarigas paraugkopas 1

wilcox.test(rokas$laba,rokas$kreisa,paired=TRUE)

## chi^2  tests 1


koki<-matrix(c(12,34,56,23,8,27,33,47,14,11),ncol=2)
rownames(koki) <- c("Priede","Egle","Berzs","Ozols","Klava")
colnames(koki) <- c("Paraug A","Paraug B")
koki

chisq.test(koki)





#Korelacijas analize

smiltaji<-read.table(file="smiltaji.txt",header=TRUE,sep="\t",dec=".")
pairs(smiltaji)

par(mfrow=c(2,2))
library(car)
for(i in 1:4){
  qqPlot(smiltaji[,i],main=names(smiltaji)[i])
}

cor(smiltaji[,1:3])

library(ltm)
rcor.test(smiltaji[,1:3])

cor.test(smiltaji$pH,smiltaji$sunas)

cor.test(smiltaji$sugas,smiltaji$smilts,method="spearman")
cor.test(smiltaji$sugas,smiltaji$smilts,method="kendall")




#Regresijas analize

bietes <- read.table(file = "bietes.txt", header = T)
str(bietes)
summary(bietes)

modelis <- lm(formula = svars ~ udens, data = bietes)
summary(modelis)

par(mfrow = c(2, 2))
plot(modelis)
par(mfrow = c(1, 1))


koef <- coefficients(modelis)
koef

udens2 <- 301
raza2 <- koef[1] + koef[2] * udens2
raza2

jaunidati <- data.frame(udens = 301)
predict(modelis, jaunidati,  interval = "prediction")



dati <- read.table(file = "renda.txt", header = T)
str(dati)
summary(dati)

modelis <- lm(formula = hron ~  dec + jan + feb + mar, data = dati)
summary(modelis)

library(car)
vif(modelis)

modelis1 <- lm(hron ~ jan + feb + mar, data = dati)
summary(modelis1)

modelis2 <- lm(hron ~ feb + mar, data = dati)
summary(modelis2)

anova(modelis,modelis2)


par(mfrow = c(2, 2))
plot(modelis2)



#Dispersijas analize

miezi <- read.table(file = "miezi.txt", header = TRUE,sep="\t",dec=".")

str(miezi)
head(miezi)

summary(miezi)

boxplot(miezi$raza~miezi$skirne)

is.factor(miezi$skirne)
miezi$skirne<-as.factor(miezi$skirne)
is.factor(miezi$skirne)

library(car)
leveneTest(y = miezi$raza, group = miezi$skirne)

anov.miezi <- aov(raza~skirne,data=miezi)
summary(anov.miezi)

275/(275+808)

TukeyHSD(anov.miezi)

library(gplots)
plotmeans(miezi$raza ~ miezi$skirne,
          connect=FALSE,xlab = "Kviešu šķirne",ylab = "Raža",
          main = "Vidējās vērtības ar 95% ticamības intervālu")


soja <- read.table(file="soja.txt",header=TRUE)

str(soja)
head(soja)

names(soja)
is.factor(soja$gaisma)
is.factor(soja$stress)

library(car)
leveneTest(y=soja$lapas,group=soja$gaisma:soja$stress)

modelis <- aov(lapas~gaisma*stress,data=soja)
summary(modelis)

kv.sum <- 42752 + 14858 + 26 + 42976
42752/kv.sum  #gaisma
14858/kv.sum  #stress

TukeyHSD(modelis,"gaisma")

kruskal.test(raza~skirne,data=miezi)