KNNRegressor is a supervised machine learning algorithm for regression. In machine learning, regression can be thought of as a mapping from one space to another where both can contain continuous values. In order to make predictions, the KNNRegressor must first be fit with an input DataSet of data points, each of which is paired (by means of a shared identifier) with another data point in an output DataSet.

Pointer

The output DataSet used with a KNNRegressor must have only 1 dimension. For an object that can perform regression with multidimensional outputs, see the MLPRegressor.

The KNNRegressor uses an internal KDTree to find an input point’s numNeighbours nearest neighbours in an input dataset. When the weight parameter equals 1, the output returned is a weighted average of those neighbours’ values from the output DataSet (this is the default). If the weight parameter is set to 0, the output returned is a simple average of the neighbours.

When training machine learning models, including the KNNRegressor, it could be important to test and validate the trained model. Learn more about this process at Training-Testing Split.

Last modified: Tue Aug 23 14 by James Bradbury
Edit File on GitHub