Accounting results modelling with neural networks: the case of an international oil and gas company

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review


Accounting results are crucial information closely monitored by managers, investors and government agencies for decision making. Understanding various endogenous and exogenous business factors affecting accounting results is an essential step in managing them. However, how to model the relationship between accounting results and their business factor antecedents remains an unresolved issue. To address this issue, this paper develops neural network (NN) models for modelling complex interactions between the business factors and accounting results. Based on empirical data from an international leading oil and gas company, 15 original data points, 8 inputs and 6 outputs are used, and 4 NN architectures in 2 training settings are tested. The experiments conducted show satisfactory results. Comparisons of various training settings suggest that a recurrent NN architecture with multiple outputs is best suited for accounting results modelling. The relative contribution factor analysis with the best-performing NN model provides new insights in understanding crucial business factors for the case company and accounting professionals to manage accounting results. As a pilot study, this paper contributes to business, accounting and finance research by providing a promising approach for accounting results modelling.

Original languageEnglish
Title of host publicationNeural Information Processing
Subtitle of host publication25th International Conference, ICONIP 2018 Siem Reap, Cambodia, December 13–16, 2018 Proceedings, Part II
EditorsLong Cheng, Andrew Chi Sing Leung, Seiichi Ozawa
Place of PublicationCham Switzerland
Number of pages11
ISBN (Electronic)9783030041793
ISBN (Print)9783030041786
Publication statusPublished - 2018
EventInternational Conference on Neural Information Processing 2018 - Siem Reap, Cambodia
Duration: 13 Dec 201816 Dec 2018
Conference number: 25th (Proceedings)

Publication series

NameLecture Notes in Computer Science
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


ConferenceInternational Conference on Neural Information Processing 2018
Abbreviated titleICONIP 2018
CitySiem Reap
Internet address


  • Accounting results
  • Business factors
  • Neural network modelling
  • Oil and gas company

Cite this