Development of an innovative framework for missing data in retail data science. (12th May 2022)

• Record Type: Journal Article
• Authors: Mahapatra, Ashok; Patnaik, Srikanta; Dash, Manoranjan; Mahapatra, Ananya
• Is Part Of: International journal of applied decision sciences. Volume 15:Number 4(2022)
• Page Start: 426
• Page End: 464
• Publication Date: 2022-05-12
• Subjects: innovative framework -- retail strategy -- retail data science -- missing value -- big data; Decision making -- Periodicals; Management science -- Periodicals; 658.403005
• Languages: English
• ISSNs: 1755-8077

Abstract

Although handling missing data and missing value imputation are widely researched subjects, missing data identification and treatment has not been pursued as a principal apparatus in retail data science applications. Critical data science derived strategies for assortment optimisation, customer purchasing behaviour and supply chain draw conclusions mostly based on the assumptions of non-missing, complete datasets. Therefore, we not only explore missing data scenarios in retail holistically: 1) from a data science perspective; 2) from an operational perspective; 3) from an implementation perspective, such that we can develop a robust framework, but also, we fill the gaps in: 1) identification; 2) treatment of missing data. To make our recommendations robust and comprehensive, we have proposed an implementable framework that harnesses the missing data scenarios in retail holistically and bridges the gaps in identification and treatment of it. At the core of the framework is a decision tree conjoining systematically derived two options trees, one from the retail industry operations and the other from the spectrum of missing data methods in the realm of data science.