Demand Estimation from Sales Transaction Data -- Practical Extensions

TL;DR

This paper addresses practical challenges in estimating demand from sales transaction data by proposing model extensions, data preprocessing techniques, and algorithms suitable for real-world data complexities.

Contribution

It introduces new algorithms and model modifications to better estimate demand from sales data with partial availability and non-homogeneous product sets.

Findings

Developed an algorithm for splitting sales data under partial availability

Extended EM algorithm for non-homogeneous product sets

Created iterative optimization algorithms for practical demand estimation

Abstract

In this paper we discuss practical limitations of the standard choice-based demand models used in the literature to estimate demand from sales transaction data. We present modifications and extensions of the models and discuss data preprocessing and solution techniques which are useful for practitioners dealing with sales transaction data. Among these, we present an algorithm to split sales transaction data observed under partial availability, we extend a popular Expectation Maximization (EM) algorithm for non-homogeneous product sets, and we develop two iterative optimization algorithms which can handle much of the extensions discussed in the paper.