One Rule

A One Rule or decision stump is a simple decision tree with a single split.

It is used as a baseline in classification tasks to compare the performance of more complex algorithms. The One Rule algorithm selects the single predictor variable that best separates the data into two classes and ignores all other variables.

For example, if you have a dataset with multiple features, One Rule will choose the feature that provides the most accurate classification and base its decision solely on that feature.

In summary, One Rule is a straightforward and interpretable algorithm that serves as a benchmark for evaluating the effectiveness of more sophisticated classification models. It is important to know because it highlights the performance of simple models and sets a baseline for comparison.