This is an in-depth tutorial written to introduce a simple, yet powerful algorithm called K-Nearest-Neighbors (KNN)
for Classification & Regression problems.
Classification
is used to predict which class a data point belongs to (discrete value)Regression
is used to predict continuous values
Our First ML model developed for Iris dataset in previous post to classify the species of iris flower is a classic example of classificaiton problem.
We will go over the intuition and theoretical detail of the algorithm, apply it to both synthetic & real-world dataset to see exactly how it works,
and gain an intrinsic understanding of its inner-workings by writing code from scratch &
tuning its most important parameter num_neighbor
to see its effect on model complexity and generalization.
Finally, we will explore the imporant parameters of the model, its strength & weakness.