Using the codes I provided below provide a comment on Cross Validation and Test & training explaining the analysis used, and why using this analysis then based on that analysis explain what should be the best model you suggest for prediction and why. And what limitations you found using this analysis .

``{r}

library(carData)

library(MASS)

library(ISLR)

library(car)

library(leaps)

```

```{r}

Salaries1<-Salaries

str(Salaries1)

attach(Salaries1)

```

### Train & test data

```{r}

set.seed(101)

# Now selecting 80% data as sample from total 'n' rows of the data

sample <- sample.int(n=nrow(Salaries1), size = floor(.80 * nrow(Salaries1)), replace = F)

train <- Salaries1[sample,]

test <- Salaries1[-sample,]

```

### Cross validation

```{r}

library(caret)

# Define training control

set.seed(123)

# Here CV stands for cross validation with 10 cross fold validation occurs

train.control<- trainControl(method = "cv", number =10)

# Train the model

model <- train(salary~., data= Salaries1, method = "lm",

trControl = train.control)

# Summarize the results

print(model)

```

```{r}

# Define training control

train.control<- trainControl(method = "LOOCV", number =10)

# Train the model

model <- train(salary~., data= Salaries1, method = "lm",

trControl = train.control)

# Summarize the results

print(model)

```

```{r}

set.seed(123)

train.control <- trainControl(method = "repeatedcv",

number = 10, repeats = 3)

# Train the model

model <- train(salary ~., data = Salaries1, method = "lm",

trControl = train.control)

# Summarize the results

print(model)

```

```{r}

model$finalModel

```

```{r}

model$resample

```

```{r}

sd(model$resample$Rsquared)

```