Confidence Score
A numerical value (typically 0-1) indicating how certain a model is about its prediction. Higher scores indicate greater confidence in the output.
Why It Matters
Confidence scores enable threshold-based decisions — accept predictions above 0.9, send those between 0.5-0.9 for human review, reject below 0.5.
Example
A document classifier assigning a 0.95 confidence score to 'invoice' (very sure) versus 0.55 to 'purchase order' (uncertain) — the low-confidence case triggers human review.
Think of it like...
Like a witness saying 'I am 95% sure it was him' versus 'I think it might have been him' — the confidence level determines how much weight to give the answer.
Related Terms
Softmax
A function that converts a vector of numbers into a probability distribution, where each value is between 0 and 1 and all values sum to 1. It is typically used as the final layer in classification models.
Classification
A type of supervised learning task where the model predicts which category or class an input belongs to. The output is a discrete label rather than a continuous value.