comp_pred

<code>comp_pred</code> provides the main wrapper for running alternative classification algorithms, such as CART (<code><a href="/link/rpart%3A%3Arpart?package=FFTrees&version=1.8.0" data-mini-rdoc="FFTrees::rpart::rpart">rpart::rpart</a></code>),
logistic regression (<code>glm</code>), support vector machines (<code>svm::svm</code>), and random forests (<code><a href="/link/randomForest%3A%3ArandomForest?package=FFTrees&version=1.8.0" data-mini-rdoc="FFTrees::randomForest::randomForest">randomForest::randomForest</a></code>).

Create, visualize, and test fast-and-frugal decision trees (FFTs) using the algorithms and methods described in Phillips, Neth, Woike & Gaissmaier (2017).
FFTs are simple and transparent decision trees for solving binary classification problems.
FFTs can be preferable to more complex algorithms because they require very little information, are easy to communicate, and are robust against overfitting.

Hansjoerg Neth

FFTrees

Generate, Visualise, and Evaluate Fast-and-Frugal Decision Trees

Nathaniel Phillips

Jan Woike

Wolfgang Gaissmaier

comp_pred function

<dl><dt>formula</dt>
<dd>A formula (usually <code>x$formula</code>, for an <code>FFTrees</code> object <code>x</code>).</dd>
<dt>data.train</dt>
<dd>A training dataset (as data frame).</dd>
<dt>data.test</dt>
<dd>A testing dataset (as data frame).</dd>
<dt>algorithm</dt>
<dd>character string. An algorithm in the set:
"lr" -- logistic regression;
"rlr" -- regularized logistic regression;
"cart" -- decision trees;
"svm" -- support vector machines;
"rf" -- random forests.</dd>
<dt>model</dt>
<dd>model. An optional existing model, applied to the test data.</dd>
<dt>sens.w</dt>
<dd>Sensitivity weight parameter (from 0 to 1, required to compute <code>wacc</code>).</dd>
<dt>new.factors</dt>
<dd>string. What should be done if new factor values are discovered in the test set?
"exclude" = exclude (i.e.; remove these cases), "base" = predict the base rate of the criterion.</dd></dl>

Arguments

<code>comp_pred</code> provides the main wrapper for running alternative classification algorithms, such as CART (<code><a href='https://rdrr.io/pkg/rpart/man/rpart.html'>rpart::rpart</a></code>),
logistic regression (<code>glm</code>), support vector machines (<code>svm::svm</code>), and random forests (<code><a href='https://rdrr.io/pkg/randomForest/man/randomForest.html'>randomForest::randomForest</a></code>).

A wrapper for competing classification algorithms. — comp_pred

<dl>

<dt>formula</dt>
<dd>A formula (usually <code>x$formula</code>, for an <code>FFTrees</code> object <code>x</code>).</dd>


<dt>data.train</dt>
<dd>A training dataset (as data frame).</dd>


<dt>data.test</dt>
<dd>A testing dataset (as data frame).</dd>


<dt>algorithm</dt>
<dd>character string. An algorithm in the set:
"lr" -- logistic regression;
"rlr" -- regularized logistic regression;
"cart" -- decision trees;
"svm" -- support vector machines;
"rf" -- random forests.</dd>


<dt>model</dt>
<dd>model. An optional existing model, applied to the test data.</dd>


<dt>sens.w</dt>
<dd>Sensitivity weight parameter (from 0 to 1, required to compute <code>wacc</code>).</dd>


<dt>new.factors</dt>
<dd>string. What should be done if new factor values are discovered in the test set?
"exclude" = exclude (i.e.; remove these cases), "base" = predict the base rate of the criterion.</dd>

</dl>

comp_pred: A wrapper for competing classification algorithms.

Description

Usage

Arguments