個案總覽

This note discusses logistic regression using binomial data. Also discussed is an explanation of how to estimate the parameters from logistic regression using Microsoft Excel (without any add-ins). The note is a follow up to Note on Logistic Regression, product #910E05, which discusses the case with Bernoulli data.

One of the most basic revenue management problems is how to allocate a fixed capacity to various types of classes of customers so that the total profit is maximized. Using the example of a flight traveling from Toronto to Vancouver, the revenue management problem is to decide, in each period, how much of the realized demand to accept and how much to reserve for the latter classes in order maximize the revenue. The crux of the problem revolves around the trade-off between spoilage and dilution. The author uses various mathematical formulae (including the static model, Littlewood's Two-Class model and n-Class model of expected marginal seat revenue) to model optimal seat-allocation outcomes.