Hands-On Prescriptive Analytics

Chapter 10. Developing Menu Priors

I will introduce several methods for developing the prior probabilities for Prescriptive Analytics in this second of two Menu Development chapters. These are used in the expected value calculations for each menu option. The expected values finalize the information decision-makers need to complete their assessments of each menu option to make a rational choice.

In this chapter, I will address these two leading questions:

1. How are beliefs incorporated in probabilities?

2. How are subjective probability weights developed?

Background

I asked an important question in the previous chapter: “Where does the menu come from?” It comes from questions decision-makers ask, regardless of their scale-view. For example, in the context of a strategic scale-view, these could originate as part of a five-year business planning process.

Although my question is of paramount importance—after all, there can be no decision without decision options (i.e., a menu)—another factor is also significantly important. If you study the use-cases, especially the airport-downtown hotel development example, you will notice probability distributions associated with the menu options. I had displayed these in Table 6-1. These distributions played an important role in the analysis of that use-case.

You should immediately ask: “What is the basis for the probabilities?” I just assumed the hotel management team, in conjunction with the DAC and urban development consultants, specified ...