Decision Taking Based on BN

Discover how to navigate scenarios, pinpoint key improvements, and improve project outcomes.

Welcome to the final lesson of this journey. We will delve into decision-making with the help of Bayesian networks. We will explore the necessity of evaluating multiple scenarios to make informed and effective decisions.

In real-world scenarios, decisions are rarely straightforward. They often involve numerous variables and uncertain circumstances, a complexity that can be effectively modeled using Bayesian networks. These networks offer many possibilities, providing various outcomes based on the input parameters. It is important to realize that these inputs are not chosen randomly but are based on the specific constraints of the context we're modeling.

Our primary objective is to understand how to navigate these multiple scenarios and make the best possible decision. To this end, we will employ a Bayesian network and observe the importance of considering context-specific constraints.

During this lesson, we will use Python and data manipulation techniques to implement these concepts. The goal is to ensure you're well-equipped to utilize these concepts in your decision-making processes, no matter the context. Stay tuned as we embark on this exciting journey of understanding and implementing decision-making models using Bayesian networks.

Two improvement scenarios to choose from

Given the objective to reduce overcost (i.e., to minimize the probability of overcost to occurs ), we are evaluating two strategies:

Individual improvement strategy: This strategy suggests improving the maturity levels of nodes one at a time. For instance, you may focus on improving the maturity level of node H from level 1 to level 5, one step at a time. The benefit of this approach is that it allows us to specifically focus on one node at a time, which could potentially reduce the risk of errors and allow for a more controlled improvement process. The next slides show the query scenarios for this strategy:

Get hands-on with 1300+ tech skills courses.