Part 3 — Simulating Scenarios
—
INTRODUCTION
Simulation is a powerful tool in the data science tool box. In this article, we’ll talk about how simulation can help us make better decisions and strategies by simulating possible scenarios. One key concept we’ll explore throughout is how we can leverage machine learning models and scenario simulation to make better decisions.
The specific topics of this article are:
- Scenario simulation for optimization
- Scenario simulation for risk management
This is the third part in multi-part series on simulation in data science. The first article covered how simulation can be used to test machine learning approaches and the second article covered using simulation to estimate the power of a designed experiment.
WHAT IS DATA SIMULATION?
The first article spends a lot more time defining simulation. To avoid redundancy, I’ll just give a quick definition here:
Data simulation is the creation of fictitious data that mimics the properties of the real-world.
Okay, with that out of the way, let’s talk about scenario simulation!
SCENARIO SIMULATION FOR OPTIMIZATION
Often, we develop machine learning models to make predictions on real data. For example, models that predict if a tumor is malignant, or if a customer is likely to default on their loan. In these cases we pass real data into our model to get predictions on real entities (patients, customers etc). When we use our machine learning models for scenario analysis, we often use simulated data to see that would happen given certain (simulated)…