F1 Skoru
The classification success officer, which takes the harmonic average of precision and sensitivity.
F1 score is an metric that takes the accuracy (precision) and sensitivity (recall) harmonic average to summarize the success of a classification model with a single number. The reason for the average use of Harmonic is that one of the two values is very low, the F1 score is also low; that is, the model can only achieve a high F1 score by negligence or simply neglecting the other. As formula, F1 = 2 × (cutting × sensitivity) / (cutting + sensitivity) is expressed in the form and receives a value between 0 and 1; approaching to 1 means both the wrong alarm and less missed case.
F1 score is particularly preferred in problems where class distribution is unbalanced, because in such cases it can be misleading to look at only accuracy (accuracy) civilization: even by predicting everything "not scam" can achieve high accuracy, but this is a model that does not work. F1 score is widely reported in areas such as medical diagnosis, spam filtering, fraud detection; in some cases, generalized versions such as F-beta score to give different weight to certainty or susceptibility.
F1 score is particularly preferred in problems where class distribution is unbalanced, because in such cases it can be misleading to look at only accuracy (accuracy) civilization: even by predicting everything "not scam" can achieve high accuracy, but this is a model that does not work. F1 score is widely reported in areas such as medical diagnosis, spam filtering, fraud detection; in some cases, generalized versions such as F-beta score to give different weight to certainty or susceptibility.