non-parametric regression

Non-parametric regression is a method used to predict or understand the relationship between variables without assuming a specific fixed mathematical form, like a straight line or a curve. Instead, it relies on the data itself to shape the prediction, adapting locally based on nearby points. This approach is flexible and effective when the true relationship is complex or unknown, allowing the model to capture intricate patterns without predefined equations. Examples include techniques like kernel smoothing and decision trees, which focus on the data's structure rather than strict formulas.