---
title: "Regression modeling "
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{regression}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---


For showing regression `SSLR` models, we will use *Airquality* dataset with 10% labeled data:

```{r eval=FALSE, warning=FALSE,message=FALSE}

library(SSLR)
library(tidymodels)
```

```{r libraries, results="hide", warning=FALSE,message=FALSE}
knitr::opts_chunk$set(
  digits = 3,
  collapse = TRUE,
  comment = "#>"
)
options(digits = 3)

library(SSLR)
library(tidymodels)
```

```{r airquality, results="hide"}
set.seed(1)

data <- airquality
#Delete column Solar.R (NAs values)
data$Solar.R <- NULL
#Train and test data
train.index  <- sample(nrow(data), round(0.7 * nrow(data)))
train <- data[ train.index,]
test  <- data[-train.index,]

cls <- which(colnames(airquality) == "Ozone")

#% LABELED
labeled.index <- sample(nrow(train), round(0.1 * nrow(train)))
train[-labeled.index,cls] <- NA
```

For example, we can train with Decision Tree:

```{r fit, results="hide"}
m <- SSLRDecisionTree(min_samples_split = round(length(labeled.index) * 0.25),
                      w = 0.3) %>% fit(Ozone ~ ., data = train)
```
Now we can use metrics from `yardstick` package:

```{r metrics}
predict(m,test)%>%
  bind_cols(test) %>%
  metrics(truth = "Ozone", estimate = .pred)
```


We can train with Random Forest:

```{r fitrf, results="hide"}
m <- SSLRRandomForest(trees = 5,  w = 0.3) %>% fit(Ozone ~ ., data = train)
```


For example, we can train with coBC:

```{r fitcobc, results="hide", eval = FALSE}
m_r <- rand_forest( mode = "regression") %>%
  set_engine("ranger")

m <- coBC(learner = m_r, max.iter = 1) %>% fit(Ozone ~ ., data = train)
```

We can train with COREG:

```{r fitcoreg, results="hide", eval = FALSE}
#Load kknn
library(kknn)
m_coreg <- COREG(max.iter = 1)  %>% fit(Ozone ~ ., data = train)
```