Regression tree model for prediction of demand with heterogeneity and censorship

• Published date: 01 April 2020
• Author: Evgeniy M. Ozhegov,Alina Ozhegova
• Date: 01 April 2020
• DOI: http://doi.org/10.1002/for.2643

Received: 6 January 2018 Revised: 28 October 2019 Accepted: 25 November 2019

DOI: 10.1002/for.2643

RESEARCH ARTICLE

Regression tree model for prediction of demand with

heterogeneity and censorship

Evgeniy M. Ozhegov1Alina Ozhegova2

1Department of Economics and Finance,

National Research University Higher

School of Economics, Perm, Russia

2Research Group for Applied Markets and

Enterprises Studies, National Research

University Higher School of Economics,

Perm, Russia

Correspondence

Evgeniy M. Ozhegov,Department of

Economics and Finance, National

Research University Higher School of

Economics, Studencheskaya St., 614000

Perm, Russia.

Email: eozhegov@hse.ru

Funding information

The Academic Fund Program at the

National Research University Higher

School of Economics (HSE), Grant/Award

Number: 18-01-0025

Abstract

In this research we analyze a new approach for prediction of demand. In the

studied market of performing arts the observed demand is limited by capacity of

the house. Then one needs to account for demand censorship to obtain unbiased

estimates of demand function parameters. The presence of consumer segments

with different purposes of going to the theater and willingness-to-pay for perfor-

mance and ticket characteristics causes a heterogeneity in theater demand. We

propose an estimator for prediction of demand that accounts for both demand

censorship and preferences heterogeneity.The estimator is based on the idea of

classification and regression trees and bagging prediction aggregation extended

for prediction of censored data. Our algorithm predicts and combines predic-

tions for both discrete and continuous parts of censored data. We show that our

estimator performs better in terms of prediction accuracy compared with esti-

mators which account either for censorship or heterogeneity only.The proposed

approach is helpful for finding product segments and optimal price setting.

KEYWORDS

censored data, demand, machine learning, performing arts, regression tree

1INTRODUCTION

Currently,firms, households, and society as a whole gener-

ate, collect, and store an enormous volume of data starting

from the level of individuals to the level of countries.

Availability of large data sets, in turn, opens the way for

modeling consumer behavior for retail, banking, and other

industries to improve their business decisions. Analysis

based on big data also matters to marketing, since cus-

tomers provide much information that may be potentially

useful for marketing decisions. Up-to-date analysis prob-

lems have been restrictedby standard econometric models.

Development of computer science techniques and large

volumes of data result in the implementation of machine

learning (ML) techniques in solving current business and

marketing problems.

In terms of marketing and business decisions, pricing is

a crucial factor for a company that aims to increase a profit.

Effective pricing strategy helps the company to determine

the price level at which it maximizes profits from the

sales of products. When the product is differentiated—that

is, possesses a set of attributes—the seller should under-

stand customers' preferences according to each attribute.

Researchers of demand analyze customers' preferences

using willingness-to-pay, which demonstrates the added

value of each attribute.

Availability of large data sets of customers' purchases

in various industries allows us to estimate demand

function. The results of estimation allow us to provide

recommendations about the significance of particular

product attributes, especially price. Preceding papers

study demand primarily on aggregated data (Lange &

Luksetich, 1984; Levy-Garboua & Montmarquette, 1996;

Schimmelpfennig, 1997; Throsby, 1994). Models of aggre-

gated data allow us to draw an inference only about

the average consumer. These conclusions may be useful

Journal of Forecasting. 2020;39:489–500. wileyonlinelibrary.com/journal/for © 2019 John Wiley & Sons, Ltd. 489