K Nearest Neighbour
-
Classifier
Voting is done, based on the majority classes of nodes within a region to determine the prediction class
-
Piecewise Linear Decision Boundary
-
K is the number of nodes / data points within a neighbour of a data point-
These K nodes are used to classify the test data point are are similar to it
-
More the value of K, the more simple the model will be
- It won’t be able to learn much
-
Less the value of K, the more complex the model will be
- It will start to memorise examples
-
-
-
Regressor
Average or Weighted Average is taken of the continuous class labels of the nodes within the region to compute the prediction class
- Can be applied on continuous values
- There exists a Piecewise constant that determines the range of each label
- The label of majority class within a region / radius is assigned to the data point