Learn R Programming

⚠️There's a newer version (2.2.10) of this package.Take me there.

emil (version 2.0.2)

Evaluation of Modeling without Information Leakage

Description

A toolbox for designing and evaluating predictive models with resampling methods. The aim of this package is to provide a simple and efficient general framework for working with any type of prediction problem, be it classification, regression or survival analysis, that is easy to extend and adapt to your specific setting. Some commonly used methods for classification, regression and survival analysis are included.

Copy Link

Version

Install

install.packages('emil')

Monthly Downloads

1

Version

2.0.2

License

GPL (>= 2)

Maintainer

Christofer Backlin

Last Published

May 20th, 2015

Functions in emil (2.0.2)

Support function for identifying missing values

Fit quadratic discriminant.

PAMR adapted dataset pre-processing

Load a package and offer to install if missing

emil_list_method

List all available methods

Wrapper for several methods to test if a variable is empty

Perform modeling

Nearest neighbors imputation

Fit linear discriminant

Dichotomize time-to-event data

fit_randomForest

Fit random forest.

Regular imputation

Fit conditional inference forest

Convert a fold to row indexes of fittdng or test set

Plots an axis the way an axis should be plotted.

get_performance

Extract prediction performance

Negative geometric mean of class specific predictive accuracy

Fit a linear model fitted with ordinary least squares

Fit GLM with LASSO, Ridge or elastic net regularization.

Basic imputation

Predict using a caret method

Get color palettes

Print a timestamped and indented log message

Prediction using nearest shrunken centroids.

Fit nearest shrunken centroids model.

Increase indentation

Replace values with something else

Plots a box around a plot

importance_pamr

Feature importance of nearest shrunken centroids.

Extending the emil framework with user-defined methods

predict_cforest

Predict with conditional inference forest

predict_randomForest

Prediction using random forest.

Introduction to the emil package

Plot Surv vector

Prediction using already trained prediction model

Extract parameter tuning statistics

Feature (variable) importance of a fitted model

Learning curve analysis

Extract predictions from modeling results

Predict the response of unknown observations

Prediction using linear model

Data preprocessing

plot.learning_curve

Plot results from learning curve analysis

is_multi_procedure

Detect if modeling results contains multiple procedures

Performance estimation functions

Extract p-value from a Cox proportional hazards model

Print a warning message if not printed earlier

Visualize resampling scheme

Fit a model using the caret package

Get names for modeling procedures

Extract p-value from a cumulative incidence estimation

Add vertical or horizontal lines to a plot

Generate resampling subschemes

Resampling schemes

Calculate ROC curves

importance_randomForest

Feature importance of random forest.

emil and dplyr integration

pvalue.survdiff

Extracts p-value from a logrank test

Extracts p-value from a competing risk model

Prediction using already trained classifier.

Predict using generalized linear model with elastic net regularization

modeling_procedure

Setup a modeling procedure

Tune parameters of modeling procedures

Extract a subset of a tree of nested lists

Extraction of p-value from a statistical test

weighted_error_rate

Weighted error rate