From input to prediction
Calculate a prediction and distinguish it from an observed outcome.
Prepared for Polaris · Original introductory content · Read & practice
Understand the idea
A model is a simplified representation of a relationship in data. A linear example uses: prediction = weight × input + bias. The input is supplied to the model; the weight controls how the output changes with it; the bias adds an offset. This calculation alone does not make a prediction reliable.
For illustration only, choose a weight of 2 and a bias of 1. Input 3 gives output 7. These numbers are not empirical findings and do not describe a real learner. Changing the input while keeping parameters fixed uses the model; it does not train it.
The idea in an example
Input 4: calculate 2 × 4 + 1 = 9. Weight and bias stay fixed. The output is an estimate under our assumptions.
Test the idea. Understand why.
Put it in your own words.
Automatically saved in this browser only; notes are not graded.