What a hyperparameter is
A hyperparameter is a setting that is chosen before a machine learning model is trained, which controls aspects of how the model learns. Unlike the values a model learns from the data during training, hyperparameters are set in advance by the people building the model. They configure the learning process itself, influencing how training proceeds and how the resulting model turns out.
Parameters versus hyperparameters
It helps to distinguish hyperparameters from ordinary parameters. Parameters are the values the model learns from the data as it trains, adjusting them to fit the examples. Hyperparameters, by contrast, are not learned from the data; they are chosen beforehand and govern the training process. This distinction is why hyperparameters are set externally while parameters emerge from learning.
Controlling how a model learns
Hyperparameters control various aspects of how a model learns, such as settings that affect the training process. Because they shape how learning happens, the choice of hyperparameters can significantly affect how well the final model performs. Good hyperparameter choices can lead to a strong model, while poor ones can hinder learning, even with the same data.
Tuning hyperparameters
Because hyperparameters strongly influence results, choosing good values for them, often called tuning, is an important part of building machine learning models. Since they are not learned automatically from the data, practitioners experiment to find values that work well. This process of adjusting hyperparameters to improve performance is a common and important activity in machine learning.
Hyperparameters in practice
In practice, setting and tuning hyperparameters is a routine part of developing machine learning models. The right choices depend on the problem, the data, and the model. Understanding hyperparameters clarifies that some important settings are decided by people before training, separate from what the model itself learns, and that these choices matter for the outcome.
Why it matters
The hyperparameter is an important concept in machine learning. Understanding what it is - a setting chosen before training that controls how a model learns - clarifies the difference between what is configured and what is learned. For anyone interested in how machine learning models are built, it is a valuable concept to understand.
Related on Skillo
See also: What is gradient descent? Explained, What is regularization in machine learning? Explained.
Sources
Published date reflects the original event date (2026-11-20). This article is original Skillo editorial written from the sources above; facts were verified in September 2026.
Leave a Reply