evaluation

EvaluateAUC()

Output:

true  positive rate: [0 0.5 0.5 1 1]
false positive rate: [0 0 0.5 0.5 1]
auc: 0.75

// Accuracy: 97%! That's pretty good. That means we were right 97% of the time
csvData := data.GetLabeledCategoricalCSVData()
EvaluateCategoricalData(csvData)

Output:

Accuracy = 0.97

csvData := data.GetLabeledCategoricalCSVData()
EvaluateCategoricalDataWithClasses(csvData)

Output:

Precision (class 0) = 1.00
Recall (class 0) = 1.00
Precision (class 1) = 0.96
Recall (class 1) = 0.94
Precision (class 2) = 0.94
Recall (class 2) = 0.96

// the MAE is 2.55 and the mean of our observed values is 14.0,
// so our MAE is about 20% of our mean value. Not very good, depending on the context.

// R-squared also gives us a general idea about the deviations of our predictions
// R-squared measures the proportion of the variance in the observed values
// that we capture in the predicted values.
// Remember that R-squared is a percentage and higher percentages are better.
// Here, we are capturing about 37% of the variance in the variable
// that we are trying to predict. Not very good.
EvaluateContinuousData(data.GetContinuousCSVData())

Output:

MAE = 2.55
MSE = 10.51
R^2 = 0.37