https://www.youtube.com/watch?v=nt63k3bfXS0&list=PLoROMvodv4rMiGQp3WXShtMGgzqpfVfbU&index=6

In this lecture, he starts off with Naive Bayes (a generative learning algorithm). To discuss this algorithm, he gives an example of an email/spam classification algorithm. A prediction on example will be as such:

Maximum likelihood:

Laplace smoothing:

The classification model may not know that the new word being enter is not spam, and it is likely that it could be placed into spam, even when it genuine email.

Multi Event model

He also starts of with Support Vector Machine. This was something I didn’t get quite well, but watched few other videos to help me somewhat on it.