REQUIRED MetricName => Str

The name of the objective metric used to measure the predictive quality of a machine learning system. This metric is optimized during training to provide the best estimate for model parameter values from data.

Here are the options:

• "MSE": The mean squared error (MSE) is the average of the squared differences between the predicted and actual values. It is used for regression. MSE values are always positive: the better a model is at predicting the actual values, the smaller the MSE value. When the data contains outliers, they tend to dominate the MSE, which might cause subpar prediction performance.
• "Accuracy": The ratio of the number of correctly classified items to the total number of (correctly and incorrectly) classified items. It is used for binary and multiclass classification. It measures how close the predicted class values are to the actual values. Accuracy values vary between zero and one: one indicates perfect accuracy and zero indicates perfect inaccuracy.
• "F1": The F1 score is the harmonic mean of the precision and recall. It is used for binary classification into classes traditionally referred to as positive and negative. Predictions are said to be true when they match their actual (correct) class and false when they do not. Precision is the ratio of the true positive predictions to all positive predictions (including the false positives) in a data set and measures the quality of the prediction when it predicts the positive class. Recall (or sensitivity) is the ratio of the true positive predictions to all actual positive instances and measures how completely a model predicts the actual class members in a data set. The standard F1 score weighs precision and recall equally. But which metric is paramount typically depends on specific aspects of a problem. F1 scores vary between zero and one: one indicates the best possible performance and zero the worst.
• "AUC": The area under the curve (AUC) metric is used to compare and evaluate binary classification by algorithms such as logistic regression that return probabilities. A threshold is needed to map the probabilities into classifications. The relevant curve is the receiver operating characteristic curve that plots the true positive rate (TPR) of predictions (or recall) against the false positive rate (FPR) as a function of the threshold value, above which a prediction is considered positive. Increasing the threshold results in fewer false positives but more false negatives. AUC is the area under this receiver operating characteristic curve and so provides an aggregated measure of the model performance across all possible classification thresholds. The AUC score can also be interpreted as the probability that a randomly selected positive data point is more likely to be predicted positive than a randomly selected negative example. AUC scores vary between zero and one: a score of one indicates perfect accuracy and a score of one half indicates that the prediction is not better than a random classifier. Values under one half predict less accurately than a random predictor. But such consistently bad predictors can simply be inverted to obtain better than random predictors.
• "F1macro": The F1macro score applies F1 scoring to multiclass classification. In this context, you have multiple classes to predict. You just calculate the precision and recall for each class as you did for the positive class in binary classification. Then, use these values to calculate the F1 score for each class and average them to obtain the F1macro score. F1macro scores vary between zero and one: one indicates the best possible performance and zero the worst.

If you do not specify a metric explicitly, the default behavior is to automatically use:

• "MSE": for regression.
• "F1": for binary classification
• "Accuracy": for multiclass classification.