2022 - Modeling Price Elasticity for Occupancy Prediction Hotel Dynamic Pricing

Table of Contents

1. Problem Definition
2. Methods
3. TODO & Questions & Further Reading

Problem Definition

1. Fliggy

Methods

1. Approach
   1. occupancy prediction
   2. revenue optimizaiton: apply a set of price, the find the max revenue
      1. Main contribution:
         1. novel demand function that explicitly models the price elasticity of demand for occupancy prediction
         2. define a parameter to represent elaciticity; define formular in the NN
      2. 3 categories of features
         1. competiveness factors
         2. temporal factors
         3. characteristic factors
2. Online experiment
   1. a baseline manual pricing strategy
3. Business insight
   1. the occupancy of luxury hotels are less price-sensitive than budget hotels as they have less substitutes in the market place
   2. For example, hot spring hotels are less price-sensitive in winter as the demand is seasonal.
   3. the features affecting 𝛽 are less dependent on price but more relevant to the room characteristics and external influences.

      TODO & Questions & Further Reading

4. Evaluation metrics -> interesting
   1. WMAPE
   2. why MAPE for classification problem
5. reference paper reading

• Previous
  2019 - Dynamic Pricing for Airline Ancillaries with Customer Context
• Next
  2023 - A Scalable and Debiased Approach to Dynamic Pricing with Causal Machine Learning and Optimization