Evaluating the model with a set threshold

The probabilities object is produced to actually evaluate the model performance with
different classification thresholds. The classification threshold can be modified for a binary
classification problem and can be used for calculating the recall score, the precision, and the
accuracy. On using a classification threshold of 0.16, these are the metrics that we get in the testing dataset:

On calculating, we get a recall score of 82.53 percent, precision of 34.02 percent, and an accuracy of 60.7 percent.