In this module, we will continue our analysis of how we go from data to decision to the focus. We'll be more on rules and why rules are important, how business rules are extracted and how they are used. So there are three basic topics in this module. First of all, to show you that rules are ubiquitous, you'll see them everywhere. The world uses rules for all kinds of things. In fact, if you think about it, there's a very nice book on Thinking, Fast and Slow, where Daniel Kahneman says that, people use rules when they have to act fast. They preprocess information and then when it comes to action, it's not much time to think. So that's why if you look at somebody who plays tennis or baseball, they practice a lot and so it's the memory which allows them to do things, but that's basically a body using rules to react to situations. Same way, in business too, there's a lot of information and then businesses extract them through actionable rules then they can go and use when implementing their processes. While I do talk a lot about rules, we're going to look at two basic ideas here. One way is how do you extract rules from data. One way is to see how similar something is to other things you have done in the past and use that to extract a rule and that's called the nearest neighbor type of algorithms. The other one is used past information to have some prior belief on what the object or what the decision is going to be, and then use information you have collected and those past information to come up with rules which are called a Bayesian methods. So these two are very classical methods of developing rules from data. So even though we operate a lot using rules, the focus of this session is when can we go from data to rules rather than just develop these rules because of what we observe without a systematic framework for analysis. I just took three examples. You see a customer walking into a store, and try to make a prediction that this customer will go to left rack and make a purchase. Trained observers can probably do that or somebody can say this person is just browsing, probably not going to buy anything. As a professor, we often try to say, "Okay, based on the performance so far, we think the student is like some of those in the past we have met who are very curious, they are able to grasp things fast, ask nice, wrote well, communicated well, work well with people. I think this student is going to do well in my class." So we jumped to perhaps some conclusions. Sometimes, more often these days, you see an unusual spending on a credit card and say, this person bought at three gas stations in a row and so no, that sounds like fraud, let's stop it and let's apply a rule. So actually, I wanted you to think of your day so far, I don't know how it's going. But think how many times you use rules in trying to answer a phone call or an email or have lunch. You will see a whole day was probably decomposed into series of decisions which depended on the way you made rules. So once you are very effect, you can say, okay, let's see why these rules came about.
