model parameter
A model parameter is a value a learning algorithm estimates from training data as it fits a model. In a neural network, for example, weights and biases are parameters; in linear regression, the fitted coefficients and intercept are parameters. Their learned values determine how the model transforms inputs into predictions. Parameters are distinct from hyperparameters, which are settings chosen outside the fitting process—such as a tree's maximum depth or a neural network's learning rate—and may be tuned using validation data. Understanding this distinction helps clarify what training learns directly and what a practitioner configures or selects when building a model.