Which term denotes the collection of values chosen before training that govern the learning process?

• A. Hyperparameters
• B. Model training parameters
• C. Epochs
• D. Confusion matrix

Answer: (A)

The main idea is that certain settings are chosen before training to shape how the model learns. These are hyperparameters. They include values like learning rate, batch size, number of layers, regularization strength, and dropout rate. They’re set prior to training and aren’t learned from the data, yet they strongly influence learning dynamics, convergence speed, and how well the model generalizes.

Model training parameters, in contrast, are the actual weights and biases that the algorithm adjusts during training. Epochs refer to how many passes the training process makes over the dataset. A confusion matrix is an evaluation tool used after training to assess classification performance. The term that denotes the collection of values chosen beforehand that govern the learning process is hyperparameters.