==
This deals with the second half of the course with the above title. The first half was taught by Prof. Naveen M. B. The second half mostly starts by looking at a glimse of what learning from data without assuming any information about the distribution generating the data; essentially searching for the pattern, if any, in the data. Let us start by looking at the following data-blue points above the hyperplane while red dots are below. Some aspects of the concept below have already been dealt with in the first half. So, it is a repetition.
-----------------------image ------------------------
Figure 1: The figure shows a simple data set that can "linearly separated" ofcourse by a hyperplane.
Although stylized, studying the above provides a lot of insights. First, note that the above can be generated randomly from a distribution but with the caveat that it can be separated by a plane. Now, we shall ask what do we mean by separating the data by a hyperplane? What is a hyperplane? Towards this, consider the following picture of a hyperplane.