Satellite big data analytics for ethical decision making in farmer’s insurance claim settlement: minimization of type-I and type-II errors

Narayan Prasad Nagendra, Gopalakrishnan Narayanamurthy*, Roger Moser

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Farmers submit claims to insurance providers when affected by sowing/planting risk, standing crop risk, post-harvest risk, and localized calamities risk. Decision making for settlement of claims submitted by farmers has been observed to comprise of type-I and type-II errors. The existence of these errors reduces confidence on agri-insurance providers and government in general as it fails to serve the needy farmers (type-I error) and sometimes serve the ineligible farmers (type-II error). The gaps in currently used underlying data, methods and timelines including anomalies in locational data used in crop sampling, inclusion of invalid data points in computation, estimation of crop yield, and determination of the total sown area create barriers in executing the indemnity payments for small and marginal farmers in India. In this paper, we present a satellite big data analytics based case study in a region in India and explain how the anomalies in the legacy processes were addressed to minimize type-I and type-II errors and thereby make ethical decisions while approving farmer claims. Our study demonstrates what big data analytics can offer to increase the ethicality of the decisions and the confidence at which the decision is made, especially when the beneficiaries of the decision are poor and powerless.
Original languageEnglish
Number of pages22
JournalAnnals of Operations Research
DOIs
Publication statusE-pub ahead of print - 24 Sep 2020

Keywords

  • Agriculture
  • Big data analytics
  • Ethical decision making
  • Farmer
  • India
  • Insurance
  • Satellite imagery

Fingerprint Dive into the research topics of 'Satellite big data analytics for ethical decision making in farmer’s insurance claim settlement: minimization of type-I and type-II errors'. Together they form a unique fingerprint.

Cite this