A statistical model is the approximation of the truth that has been captured through data and mathematics or statistics, and acts as an enabler here. This approximation is used to predict an event. A statistical model is nothing but a mathematical equation. 

For example, let's say we reach out to a bank for a home loan. What does the bank ask us? The first thing they would ask us to do is furnish lots of documents such as salary slips, identity proof documents, documents regarding the house we are going to purchase, a utility bill, the number of current loans we have, the number of dependants we have, and so on. All of these documents are nothing but the data that the bank would use to assess and check our creditworthiness:

What this means is that your creditworthiness is a function of the salary, number of loans, number of dependants, and so on. We can arrive at this equation or relationship mathematically.

In the next section, we will see how models learn and how the model can keep getting better.