The precision is also called the positive predicted value, and it is the ratio: TP / (TP + FP) where TP is the number of true positives, and FP is the number of false positives. The precision (intuitively) is the ability of the classifier of not labeling as positive a sample that is negative. The f1-score is also known as the balanced f-score or the f-measure. The f1-score can be interpreted as a weighted average of the precision and recall, where the f1-score best value is 1, and the worst is 0. The relative contributions of precision and recall to the f1-score are equal. The formula for the f1-score is shown in Eq. 2 :
f1-score =2 * (precision * recall) / (precision + recall). (2)
The f1-score in multi-class and multi-label cases is the average of the f1-score for each class, with weighting that depends on the average parameter. The support is the number of records used, i.e., the numbers of spectra for each class. The accuracy classification score in multi-label classification computes the accuracy subset: the set of labels predicted for a sample must exactly match the corresponding set of labels in y_true. The reported averages include the macro average (averaging the unweighted mean for each label) and the weighted average (averaging the support-weighted mean for each label).
Other parameters for the model evaluations were obtained using areas under the receiver operator curve (ROC) plots. This was generated using the ML method selected to evaluate the “sensor” performance. The ROC plots allow inspection of the fundamental trade-off in the models between TP and FP. This provides much more information than a straightforward accuracy calculation. When comparing two models, the ROC plots clearly show that a curve that is entirely over another represents a model with better results regardless of the threshold used. The area under a ROC plot is equal to the probability that a randomly selected positive case will receive a higher score than a randomly chosen negative case. In other words, it is the probability of sensing the HE.