Machine learning parameters are values learned from training data to minimize prediction errors. For example, in a uniform distribution for bus arrival times, parameters $latex a$ and $latex b$ define the range. They are the model’s knobs for accurate predictions.