So far, we have discussed how to precisely define the data. Now it's time to discuss how to precisely define the problem itself.
The first thing to notice is that modeling is a very broad topic and modeling techniques vary significantly from problem to problem.
In machine learning, for example, people usually classify problems according to the modeling technique: Is it a supervised or an unsupervised learning problem? If it's supervised, is it a regression or a classification problem?
In operations research, problems can be modeled using simulation, linear programming, or constraint programming, for instance. Again, the right choice highly depends on the problem and available resources to solve the problem.
Choosing the right modeling technique for a given problem requires skills that you will need to develop. Of course, it's insane to think that someone would be able to learn how to model with all the available techniques. In addition, with the choice of a modeling technique, we are implicitly converging towards a small subset of algorithms which we also need to dominate in order to solve the problem.
Given that there are so many modeling techniques, you need to prioritize what to focus on. For that, make sure to consider the types of problems that you are most interested in and your current skills.
In addition, it's important to keep a close eye on the trends in the market. For example, at the time we wrote this, there were many modeling processes in machine learning that were being completely automated.
Similarly, in the operations research world, dynamic programming has been largely replaced with mixed-integer programming techniques because the latter is more representative and optimization solvers have become incredibly more powerful than they used to be in the 80s and 90s.
We know that this information doesn't help you much in defining your priorities, but we hope that it gives you some awareness to make a more informed decision. Also, we strongly encourage you to have a chat with your Mip Mentor now. He will definitely help you design a path that makes sense for you.
Many times throughout your career, you will find yourself in a situation where you don't know how to model and solve a particular problem. It happens to all of us. At that time, you will see how invaluable it is to be supported by a community like Mip Wise. In fact, support is the only thing we can guarantee, really. Everything else is a challenge and opportunity to innovate!
In the next section, you will learn about algorithms.