RStudios

Question 1

One way anova

Answer 1

mean(data_set$X)
sd(data_set$X)

fit <- aov(independent variable ~ dependent variable, data = data_set)

summary(fit)

Question 2

Two way anova

Answer 2

insects_data <- read.csv(‘insects_data.csv’)

mean(insects_data$counts[insects_data$species == ‘Megacrania’ & insects_data$season == ‘Spring’])

fit <- aov(counts ~ species + season + species*season, insects_data)

summary(fit)

Question 3

RBCD anova

Answer 3

data <- read.csv(‘oat_variety.csv’)

mean(data$yield[data$variety == ‘Golden.rain’])
var(data$yield[data$variety == ‘Golden.rain’])

fit <- aov(yield ~ variety + plot, data)

summary(fit)

Question 4

SLR

Answer 4

library(readxl)

data <- read_excel(‘calls.xlsx’)

mean(data$Calls)
var(data$Calls)
sd(data$Calls)

cor( y = data$Executions, x = data$Calls)
plot( x = data$Calls, y = data$Executions, pch = 20, xlab = “Incoming phone calls”, ylab = “Trade executions per day”)

fit <- lm(Executions ~ Calls, data)

summary(fit)
anova(fit)
confint(fit)

Question 5

MLR

Answer 5

data <- read.csv(‘multreg.csv’)
View(data)

cor(data[,0:5])
cor(data[,c(“Price”, “PlotSize”, “FloorArea”, “Trees”, “Distance”)])

str(data)
data$Pool <- as.factor(data$Pool)
str(data)

fit <- lm(Price ~ PlotSize + FloorArea + Trees + Distance + Pool + PlotSize*FloorArea, data)

summary(fit)
confint(fit)
predict(fit)

Question 6

Model Building & Logistic Regression

Answer 6

logreg_data <- read.csv(‘logreg.csv’)
fit <- glm(cases ~ sex + income, data = logreg_data, family = ‘binomial’)
summary(fit)

mb_data <- read.csv(‘step.csv’)
str(mb_data)
mb_data$medschl <- as.factor(mb_data$medschl)
mb_data$region <- as.factor(mb_data$region)

fit.full <- lm(length ~ . , mb_data)
fit.empty <- lm(length ~ 1, mb_data)
step.model <- step(fit.empty, scope = formula(fit.full), direction = ‘forward’)

Question 7

Time Series 1

Answer 7

library(forecast)
library(fpp3)
library(ggplot2)

df <- forecast::wineind
autoplot(df) + ggtitle(‘Australian total wine sales’) + ylab(“litres”)

(sum(df[61:72])/12 + sum(df[62:73])/12)/2
autoplot(df) + autolayer(ma(df,12), series = “MA(12)”)

decomp <- decompose(df, type = ‘multiplicative’)
plot(decomp)

simp_mod <- snaive(df, h = 24)
autoplot(df) + autolayer(simp_mod)

Question 8

Time Series 2

Answer 8

dt <- window(df,start = c(1980,1),end = c(1985,12))

lam <- BoxCox.lambda(dt)
dt_trans1 <- BoxCox(dt,lambda = lam)
dt_trans2 <- diff(dt_trans1,lag = 12)
autoplot(dt_trans2)
ndiffs(dt_trans2)

Acf(dt_trans2)
checkresiduals(dt_trans2)

Question 9

Time Series 3

Answer 9

hw_mod <- hw(train, seasonal = ‘multiplicative’, h = 24)
autoplot(train) + autolayer(hw_mod)

hw_mod[[“model”]]
ggAcf(train)

accuracy(hw_mod,test)

Question 10

Time Series 4

Answer 10

dt <- ts(fpp3::us_change$Consumption)

train <- window(dt,end = 170)
test <- window(dt,start = 171)

ggAcf(train)
ggPacf(train)

mod1 <- Arima(dt, order = c(3,0,3))
mod2 <- auto.arima(dt)

checkresiduals(mod2)