Example: Credit Card Applications
For the next couple of sections, we are going to follow an example to learn machine learning concepts and to see how to use a package called sklearn.
Previously we talked about classification problems. A common classification problem that occurs for banks daily is the following: Should we give (insert person name here) a credit card? Usually the bank will decide this based on the person's credit score and income level. Let's build a system to automate this.
The first thing we are going to need is existing data on whether someone gets a credit card. Let's imagine the bank gave us the following applications and their results:
| Credit Score | Income ($) | Approved? |
|---|---|---|
| 700 | 80,000 | Yes |
| 650 | 50,000 | Yes |
| 800 | 50,000 | Yes |
| 550 | 40,000 | No |
| 570 | 70,000 | No |
| 500 | 75,000 | No |
| 840 | 90,000 | Yes |
Great. Continue to the next page to see how to build a classifier for this data!