Supervised Learning

In supervised learning, we are given a data set and already know what our correct output should look like, having the idea that there is a relationship between the input and the output.

---

Let's see an example about Housing price prediction:

 

Housing price prediction

This data is prices of house sold by house size. This values are continuous. So we can use data to create a linear function or quadratic function etc, for prediction. If we have a size of house, we can predict how much the housing price is. This case is about regression problem.

 

The other case is about classification probelm.

 

Breast cancer (malignant, benign)

This data is Malignant diagnosis data according to tumor size. This is consist of discrete value outputs (0 or 1) instead of continuous values. we can also predict consequence using this data. It turns out that in classification problems, sometimes you can have more than two possible values for the output.

 

When we have additional "Age" data of this case, the graph can be changed like below:

 

Breast cancer (malignant, benign)

In this case, we can divied into two like the green line. In other words, the data is classified by this line.

 

In some learning problems, not only three features are given, but infinite number of features an attributes may be given. Computer memory is not infinite. So, we need to learn how to handle this data.

---

Supervised learning problem are categorized into "regression" and "classification" problems. In a regression problem, we are trying to predice results within a continuous output, meaning that we are trying to map input variables to some continuous function. In a classification problem, we are instead trying to predict results in a discrete output. In other words, we are trying to map input variables into discrete categories.

---

Q. You’re running a company, and you want to develop learning algorithms to address each of two problems.

Problem 1:

You have a large inventory of identical items. You want to predict how many of these items will sell over the next 3 months.

Problem 2:

You’d like software to examine individual customer accounts, and for each account decide if it has been hacked/compromised.

​

Should you treat these as classification or as regression problems?

​

Ans. Problem 1 is about regression problem and Problem 2 is about classification problem.