Skip to content

What is Hyperparameter?

A hyperparameter is a value or setting chosen by the user before training a machine learning model that controls the learning process itself.

Unlike model parameters, which are learned automatically from the data during training, hyperparameters must be set in advance and are not updated by the algorithm.

They influence how quickly or effectively the model learns, how complex the model can become, and how it avoids problems like overfitting.

Common ways to select good hyperparameters include manual tuning, grid search, random search, or automated methods like Bayesian optimization.

Example

When training a neural network, you might set the learning rate to 0.01 and the number of hidden layers to 3; these choices are hyperparameters that stay fixed while the model weights are learned from data.

Why it matters

Hyperparameters can dramatically affect a model's accuracy and generalization, so choosing them well is essential for building effective AI systems today.

Frequently asked questions

Hyperparameters are set before training and control the learning process, while model parameters are learned from the data during training.