install.packages("FSelectorRcpp")

devtools::install_github("mi2-warsaw/FSelectorRcpp")

# windows users should have Rtools for devtools installation
# https://cran.r-project.org/bin/windows/Rtools/

info_gain <-
  information_gain(             # Calculate the score for each attribute
    formula = Species ~ .,      # that is on the right side of the formula.
    data = iris,                # Attributes must exist in the passed data.
    type  = "infogain",         # Choose the type of a score to be calculated.
    threads = 2                 # Set number of threads in a parallel backend.
  )

info_gain
    attributes importance
1 Sepal.Length  0.4521286
2  Sepal.Width  0.2672750
3 Petal.Length  0.9402853
4  Petal.Width  0.9554360

cutted_attrs <-
  cut_attrs(                    # Then take attributes with the highest rank.
    attrs = info_gain,
    k = 2                       # For example: 2 attrs with the higehst rank.
  )

cutted_attrs
[1] "Petal.Width"  "Petal.Length"

new_formula <-
  to_formula(                   # Create a new formula object with 
    attrs = cutted_attrs,       # the most influencial attrs.
    class = "Species"           
  )

new_formula
Species ~ Petal.Width + Petal.Length
<environment: 0x29e3080>

model <-
  glm(
    formula = new_formula,      # Use that formula in any classification algorithm.
    data = iris,                
    family = "binomial"         
  )

model

Call:  glm(formula = new_formula, family = "binomial", data = iris)

Coefficients:
 (Intercept)   Petal.Width  Petal.Length  
      -69.45         33.89         17.60  

Degrees of Freedom: 149 Total (i.e. Null);  147 Residual
Null Deviance:      191 
Residual Deviance: 5.17e-09     AIC: 6

information_gain(
    formula = Species ~ .,
    data = iris,
    type  = "infogain",
    threads = 2
  ) %>%
  cut_attrs(
    k = 2
  ) %>%
  to_formula(
    attrs = .,
    class = "Species"
  ) %>%
  glm(
    formula = .,
    data = iris,
    family = "binomial"
  )

eval_adjR2_lm <-      # Create a scorer function.
  function(
    attributes,       # That takes the currently considered subset of attributes
    data,             # from a specified dataset.
    dependent = 
      names(data)[1]  # To find features that best describe the dependent variable.
  ) {
    summary(          # In this situation we take the adj.r.squared statistic
      lm(             # from the summary of a linear model object.
        to_formula(   # This is the score to use to choose between considered 
          attributes, # subsets of attributes.
          dependent
        ),
        data = data)
    )$adj.r.squared
  }

set.seed(42)
random <- sample(x = letters, size = 150, replace = TRUE)
ran_iris <- cbind(iris, random)
head(ran_iris)
  Sepal.Length Sepal.Width Petal.Length Petal.Width Species random
1          5.1         3.5          1.4         0.2  setosa      x
2          4.9         3.0          1.4         0.2  setosa      y
3          4.7         3.2          1.3         0.2  setosa      h
4          4.6         3.1          1.5         0.2  setosa      v
5          5.0         3.6          1.4         0.2  setosa      q
6          5.4         3.9          1.7         0.4  setosa      n

fs <- feature_search(    # works in 2 modes – 'exhaustive' and 'greedy'
  attributes =
    names(ran_iris)[-1], # takes attribues and creates combinations of it's subsets
  fun = eval_adjR2_lm,   # calculates the score of a subset that depends on the
  data = ran_iris,       # evaluator function passed in the 'fun' parameter
  mode = "exhaustive",   # 'exhaustive' means to check all possible 
  sizes =                # attributes' subset combinations 
    3:5                  # of sizes passed in 'sizes'.
)

fs$fun
function(
    attributes,       # That takes the currently considered subset of attributes
    data,             # from a specified dataset.
    dependent = 
      names(data)[1]  # To find features that best describe the dependent variable.
  ) {
    summary(          # In this situation we take the adj.r.squared statistic
      lm(             # from the summary of a linear model object.
        to_formula(   # This is the score to use to choose between considered 
          attributes, # subsets of attributes.
          dependent
        ),
        data = data)
    )$adj.r.squared
  }
<bytecode: 0x39d3a10>

fs$call
feature_search(attributes = names(ran_iris)[-1], fun = eval_adjR2_lm, 
    data = ran_iris, mode = "exhaustive", sizes = 3:5)