May 2019
Intermediate to advanced
162 pages
4h 24m
English
A classic KNN algorithm will compute the distances between the training samples and store them in a distance-partitioned heap structure, such as KDTree or a ball tree, which are essentially sorted-heap binary trees. We then query the tree for test samples. Our approach in this class is going to be a little bit different in order to be a bit more intuitive and readable.
In the next section, we'll cover how we can implement it from scratch.
Read now
Unlock full access