Accuracy
Accuracy is our most commonly used evaluation index, which is the proportion of correct predictions in all instances, but when there is an imbalance in the data, the accuracy cannot fully evaluate the performance of the model.
Accuracy
The accuracy rate indicates the proportion of the actual examples that are classified as positive examples.
Recall rate
Confusion matrix
True Positive (TP): predict the number of positive classes as positive classes
True Negative (TN): predict the number of negative classes as negative classes
False Positive (FP): The number of negative classes predicted as positive classes (Type I error)
False Negative (FN): The number of positive classes predicted as negative classes (Type II error)
Model tuning parameters
Cross-validation
In order to allow the data to be verified and trained
Training data (training + verification)
K-fold on cross-validation
Grid search
Each parameter will have a viewing effect, and select the parameters with good effects
Combination of parameters