What do you call the parameters you set in the model before training rather than the ones learned by the algorithm?

What do you call the parameters you set in the model before training rather than the ones learned by the algorithm?

• A. Coefficients
• B. Learning Parameters
• C. Hyperparameters
• D. Default Settings

Answer: (C)

Hyperparameters are the settings you configure before training to control how the model learns. They shape the learning process and model capacity, yet they aren’t updated during training. Examples include learning rate, number of layers, regularization strength, and batch size. The algorithm then learns the actual model parameters—weights and biases—from the data. That separation is why hyperparameters is the best label: they are the pre-set knobs you tune, while the learned parameters are what the training process adjusts.

Hyperparameters are the settings you configure before training to control how the model learns. They shape the learning process and model capacity, yet they aren’t updated during training. Examples include learning rate, number of layers, regularization strength, and batch size. The algorithm then learns the actual model parameters—weights and biases—from the data. That separation is why hyperparameters is the best label: they are the pre-set knobs you tune, while the learned parameters are what the training process adjusts.