measures: Performance measures.
Description
A performance measure is evaluated after a single train/predict step and returns a single number to assess the quality
of the prediction (or maybe only the model, think AIC).
The measure itself knows whether it wants to be minimized or maximized and for what tasks it is applicable.
See below for a list of already implemented measures.
If you want a measure for a misclassification cost matrix, look at makeCostMeasure.
If you want to implement your own measure, look at makeMeasure.Classification:
{
Mean misclassification error.}
- acc
{
Accuracy.}
- ber
{
Balanced error rate. Mean of misclassification error rates on all individual classes.}
- tp
{
True positives.}
- tpr
{
True positive rate, also called hit rate or recall.}
- fp
{
False positives, also called false alarms.}
- fpr
{
False positive rate, also called false alarm rate or fall-out.}
- tn
{
True negatives, also called correct rejections.}
- tnr
{
True negative rate. Also called specificity.}
- fn
{
False negatives, also called misses.}
- fnr
{
False negative rate.}
- ppv
{
Positive predictive value, also called precision.}
- npv
{
Negative predictive value.}
- fdr
{
False discovery rate.}
- f1
{
F1 measure.}
- mcc
{
Matthews correlation coefficient.}
- gmean
{
G-mean, geometric mean of recall and specificity.}
- gpr
{
Geometric mean of precision and recall.}
- auc
{
Area under the curve.}
- multiclass.auc
{
Area under the curve for multiclass problems. Calls pROC::multiclass.roc.}Usage
nonenone
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
meancosts
mcp
format
nonecode
bercr
- Sum of squared errors
- Mean of squared errors
- Median of squared errors
- Sum of absolute errors
- Mean of absolute errors
- Median of absolute errors
- Root mean square error
- Concordance index
- Mean costs of the predicted choices.
- Misclassification penalty, i.e. average difference between
costs of oracle and model prediction.
- Time of fitting the model
- Time of predicting test set
- timetrain + trainpredict
- Percentage of original features used for model, useful for feature selection.
item
- mse
- medse
- sae
- mae
- medae
- rmse
- mcp
- timepredict
- timeboth
- featperc