I've trained a classifier model using Vowpal Wabbit to decide if a person is Male or Female based on name alone. I assigned labels Male=0 and Female=1. When I ran Vowpal Wabbit in prediction mode, the output had values varying between 0.0 and 1.0. I interpreted this as p <= 0.5 ==> Male and p > 0.5 ==> Female. With this the accuracy of prediction was 85%.
I'm now trying to train a classifier model for another problem where there are 200 labels instead of just two like above. The labels go from 0 to 199.
The prediction output has values between 0.0 and 199.0. When I get a value such as 99.56 (for example), how do I interpret this? Does it map to label 100?