Unified Interface for Machine Learning Models

Description

Provides a unified R6-based interface for various machine learning models with automatic interface detection, consistent cross-validation, model interpretations via numerical derivatives, and visualization. Supports both regression and classification tasks with any model function that follows R's standard modeling conventions (formula or matrix interface).

Package Content

Index of help topics:

Model                   Unified Machine Learning Interface using R6
benchmark               Benchmark Multiple Models with Cross-Validation
                        and Model-Specific Parameters
cross_val_score         Cross-Validation for Model Objects
extract_probabilities   Extract probability predictions from any R
                        model in a standardised format
formula_to_matrix       Convert a formula-based model to a matrix
                        interface
matrix_to_formula       Convert a matrix-based model to a formula
                        interface
train_test_split        Split data into training and test sets
unifiedml-package       Unified Interface for Machine Learning Models

Maintainer

T. Moudiki <thierry.moudiki@gmail.com>

Author(s)

T. Moudiki [aut, cre]

Unified Machine Learning Interface using R6

Description

Provides a consistent interface for various machine learning models in R, with automatic detection of formula vs matrix interfaces, built-in cross-validation, model interpretability, and visualization.

An R6 class that provides a unified interface for regression and classification models with automatic interface detection, cross-validation, and interpretability features. The task type (regression vs classification) is automatically detected from the response variable type.

Public fields

Methods

Public methods

• Model$new()

• Model$fit()

• Model$predict()

• Model$predict_proba()

• Model$print()

• Model$summary()

• Model$plot()

• Model$clone_model()

• Model$clone()

Method new()

Initialize a new Model

Usage

Model$new(model_fn)

Arguments

Returns

A new Model object

Method fit()

Fit the model to training data

Automatically detects task type (regression vs classification) based on the type of the response variable y. Numeric y -> regression, factor y -> classification.

Usage

Model$fit(X, y, ...)

Arguments

Returns

self (invisible) for method chaining

Method predict()

Generate predictions from fitted model

Usage

Model$predict(X, ...)

Arguments

Returns

Vector of predictions

Method predict_proba()

Predict probabilities from fitted model

Usage

Model$predict_proba(X)

Arguments

Method print()

Print model information

Usage

Model$print()

Returns

self (invisible) for method chaining

Method summary()

Compute numerical derivatives and statistical significance

Uses finite differences to compute approximate partial derivatives for each feature, providing model-agnostic interpretability.

Usage

Model$summary(h = 0.01, alpha = 0.05)

Arguments

Details

The method computes numerical derivatives using central differences.

Statistical significance is assessed using t-tests on the derivative estimates across samples.

Returns

A data.frame with derivative statistics (invisible)

Method plot()

Create partial dependence plot for a feature

Visualizes the relationship between a feature and the predicted outcome while holding other features at their mean values.

Usage

Model$plot(feature = 1, n_points = 100)

Arguments

Returns

self (invisible) for method chaining

Method clone_model()

Create a deep copy of the model

Useful for cross-validation and parallel processing where multiple independent model instances are needed.

Usage

Model$clone_model()

Returns

A new Model object with same configuration

Method clone()

The objects of this class are cloneable with this method.

Usage

Model$clone(deep = FALSE)

Arguments

Author(s)

Your Name

Examples

# Regression example with glmnet
library(glmnet)
X <- matrix(rnorm(100), ncol = 4)
y <- 2*X[,1] - 1.5*X[,2] + rnorm(25)  # numeric -> regression

mod <- Model$new(glmnet::glmnet)
mod$fit(X, y, alpha = 0, lambda = 0.1)
mod$summary()
predictions <- mod$predict(X)

# Classification example  
data(iris)
iris_binary <- iris[iris$Species %in% c("setosa", "versicolor"), ]
X_class <- as.matrix(iris_binary[, 1:4])
y_class <- droplevels(iris_binary$Species)  # factor -> classification

mod2 <- Model$new(e1071::svm)
mod2$fit(X_class, y_class, kernel = "radial")
predictions <- mod2$predict(X_class)
mod2$predict_proba(X_class)

Benchmark Multiple Models with Cross-Validation and Model-Specific Parameters

Description

Perform k-fold cross-validation on a list of models, using model-specific parameters. Supports verbose messages and a progress bar.

Usage

benchmark(
  models,
  X,
  y,
  cv = 5L,
  scoring = NULL,
  params = NULL,
  cl = NULL,
  show_progress = FALSE,
  verbose = TRUE
)

Arguments

Value

A list containing the CV scores for each model.

Examples

## Not run: 
library(randomForest)

X <- iris[, 1:4]
y <- iris$Species

models <- list(
  glm  = Model$new(caret::train),
  rf   = Model$new(randomForest::randomForest),
  xgb  = Model$new(caret::train)
)

params <- list(
  glm = list(method = "glmnet",
             tuneGrid = data.frame(alpha = 0, lambda = 0.01),
             trControl = trainControl(method = "none")),
  rf  = list(ntree = 150),
  xgb = list(method = "xgbTree",
             tuneGrid = data.frame(nrounds = 150, max_depth = 3, eta = 0.3,
                                   gamma = 0, colsample_bytree = 1,
                                   min_child_weight = 1, subsample = 1),
             trControl = trainControl(method = "none"))
)

results <- benchmark(models, X, y, cv = 5, params = params,
                     show_progress = TRUE, verbose = TRUE)
print(results)

## End(Not run)

Cross-Validation for Model Objects

Description

Perform k-fold cross-validation with consistent scoring metrics across different model types. The scoring metric is automatically selected based on the detected task type.

Usage

cross_val_score(
  model,
  X,
  y,
  cv = 5,
  scoring = NULL,
  show_progress = TRUE,
  verbose = TRUE,
  cl = NULL,
  seed = 123,
  fit_params = NULL,
  predict_params = NULL
)

Arguments

Value

Vector of cross-validation scores for each fold

Examples

## Not run: 
library(glmnet)
X <- matrix(rnorm(100), ncol = 4)
y <- 2*X[,1] - 1.5*X[,2] + rnorm(25)  # numeric -> regression

mod <- Model$new(glmnet::glmnet)
(cv_scores <- cross_val_score(mod, X, y, cv = 5))  # auto-uses RMSE
mean(cv_scores)  # Average RMSE

cross_val_score(mod, X, y,
fit_params     = list(alpha = 0, lambda = 0.1),
predict_params = list(type = "response"))

cross_val_score(mod, X, y,
fit_params     = list(alpha = 0.5, lambda = 0.1),
predict_params = list(type = "response"))

# Custom scoring: R-squared
r2 <- function(true, pred) {
  ss_res <- sum((true - pred)^2)
  ss_tot <- sum((true - mean(true))^2)
  1 - ss_res / ss_tot
}

(cv_scores4 <- cross_val_score(mod, X, y, cv = 5, scoring = r2))
mean(cv_scores4)  # Average R²

# Classification with accuracy scoring
data(iris)
X_class <- iris[, 1:4]
y_class <- iris$Species  # factor -> classification
mod2 <- Model$new(e1071::svm)
(cv_scores2 <- cross_val_score(mod2, X_class, y_class, cv = 5))  # auto-uses accuracy
mean(cv_scores2)  # Average accuracy

iris_bin <- iris[iris$Species != "virginica", ]
X_bin <- iris_bin[, 1:4]
y_bin <- droplevels(iris_bin$Species)
(cv_scores3 <- cross_val_score(mod2, X_bin, y_bin, cv = 3, 
scoring="f1", fit_params=list(kernel="polynomial")))  
mean(cv_scores3)  # Average F1

## End(Not run)

Extract probability predictions from any R model in a standardised format

Description

Extract probability predictions from any R model in a standardised format

Usage

extract_probabilities(model, X, y_train = NULL, verbose = FALSE)

Arguments

Value

Numeric matrix of shape n_samples x n_classes with column names as class labels. Attributes extraction_method and model_class record how predictions were obtained.

Convert a formula-based model to a matrix interface

Description

Wraps a model function that expects formula + data so it can be called with a plain X (data.frame or matrix) and y (vector). Factors in X are preserved; special column names are safely backtick-quoted in the generated formula so they survive the formula parser.

Usage

formula_to_matrix(fit_func, predict_func = stats::predict)

Arguments

Value

A named list with two elements:

fit(X, y, weights, ...)

Fits the model. X is a data.frame (or coercible matrix), y is the response vector. Extra arguments are forwarded to fit_func.

predict(model, newdata, ...)

Generates predictions. newdata must have the same columns as the X used in fit. Extra arguments are forwarded to predict_func.

Examples

lm_matrix <- formula_to_matrix(lm)
X <- data.frame(wt = mtcars$wt, hp = mtcars$hp, cyl = factor(mtcars$cyl))
y <- mtcars$mpg
model <- lm_matrix$fit(X, y)
lm_matrix$predict(model, X[1:5, ])

Convert a matrix-based model to a formula interface

Description

Wraps a model function that expects a numeric matrix X and a response vector y (like glmnet::glmnet) so it can be called with the familiar formula + data interface. The formula is expanded via model.matrix, which handles factor dummy-coding, interactions, and inline transformations automatically.

Usage

matrix_to_formula(
  fit_func,
  predict_func = function(model, newX, ...) stats::predict(model, newdata = newX, ...)
)

Arguments

Value

A named list with two elements:

fit(formula, data, ...)

Fits the model. The formula is expanded with model.matrix; the intercept column is dropped before passing to fit_func (add it back via ... if your model needs it). Extra arguments are forwarded to fit_func.

predict(model, newdata, ...)

Generates predictions. newdata is expanded with the same model.matrix terms captured at fit time. Extra arguments are forwarded to predict_func.

Examples

## Not run: 
glmnet_formula <- matrix_to_formula(
  fit_func = glmnet::glmnet,
  predict_func = function(model, newX, ...) {
    glmnet::predict.glmnet(model, newx = newX, s = 0.01, ...)
  }
)
model <- glmnet_formula$fit(mpg ~ wt + hp + factor(cyl), data = mtcars)
glmnet_formula$predict(model, newdata = mtcars[1:5, ])

## End(Not run)

Split data into training and test sets

Description

Randomly splits a feature matrix or data.frame and its corresponding response vector into training and test subsets.

Usage

train_test_split(X, y, test_size = 0.2, seed = NULL)

Arguments

Value

A named list with four elements:

Examples

# matrix input
X <- iris[, 1:4]
y <- iris$Species
d <- unifiedml::train_test_split(X, y, test_size = 0.3, seed = 42)
dim(d$X_train)  # 105 x 4
dim(d$X_test)   #  45 x 4

# data.frame input
d2 <- unifiedml::train_test_split(iris[, 1:4], iris$Species, test_size = 0.2)
is.data.frame(d2$X_train)  # TRUE