1st exercise
Binary Classification:
Hypothesis space can be considered as any space which divide the data in two sets. As it was shown in examples during the lecture, the line can separate the date in two groups: above and under the line, inside and outside the figure.
Multy-class Classification(p classes):
It is almost the same case as for binary classification, but only difference is that the space can be divided in more than two groups (classes).
Regression:
With regression we not making a goal to split data into groups, but to find a mapping function based on data.
It can be considered as a set of lines (linear regression) or other functions of a higher degree of one, like what we did in the polynomial regression in the last session. As it was shown in the class, we need to initialize the rule before trying to classify the space.
2nd exercise