This shows training and predictive experiment for using threshold different than 0.5.
In this example, we will use threshold as 0.6 instead of 0.5 for classifying income > 50K. To do this, we will drop "Scored Labels", add math operation to check for "scored probability" > 0.6, and then add this column to main scored data set. Finally, we will rename column to be "Scored Labels" (Optional) ![enter image description here] : http://neerajkh.blob.core.windows.net/images/ThresholdDifferentThan0.5Capture.PNG