Deep learning in business analytics and operations research: Models, applications and managerial implications (1806.10897v3)

Abstract: Business analytics refers to methods and practices that create value through data for individuals, firms, and organizations. This field is currently experiencing a radical shift due to the advent of deep learning: deep neural networks promise improvements in prediction performance as compared to models from traditional machine learning. However, our research into the existing body of literature reveals a scarcity of research works utilizing deep learning in our discipline. Accordingly, the objectives of this overview article are as follows: (1) we review research on deep learning for business analytics from an operational point of view. (2) We motivate why researchers and practitioners from business analytics should utilize deep neural networks and review potential use cases, necessary requirements, and benefits. (3) We investigate the added value to operations research in different case studies with real data from entrepreneurial undertakings. All such cases demonstrate improvements in operational performance over traditional machine learning and thus direct value gains. (4) We provide guidelines and implications for researchers, managers and practitioners in operations research who want to advance their capabilities for business analytics with regard to deep learning. (5) Our computational experiments find that default, out-of-the-box architectures are often suboptimal and thus highlight the value of customized architectures by proposing a novel deep-embedded network.

• The paper demonstrates that deep neural networks outperform traditional methods in capturing complex, non-linear business data.
• Case studies reveal significant improvements in prediction accuracy and operational efficiency in areas like sales and risk management.
• The paper introduces a novel deep-embedded network architecture that integrates categorical data effectively, yielding quantifiable economic benefits.

Deep Learning in Business Analytics and Operations Research: An Analytical Synthesis

The paper by Kraus, Feuerriegel, and Oztekin provides a comprehensive examination of the role of deep learning in business analytics and operations research, detailing the models, applications, and implications for managerial practice. The authors focus on the potential of deep neural networks (DNNs) to exceed traditional machine learning methods in predictive performance, due primarily to their ability to capture complex, non-linear relationships in data. Despite its potential, they note a relative scarcity of research and practical application of deep learning within the field, highlighting the necessity for further exploration and adoption.

Key Findings and Contributions

1. Review of Existing Literature: The paper identifies a limited number of studies applying deep learning to business analytics within leading operations research journals. This highlights a gap and opportunity for the field to more broadly incorporate deep learning methodologies.
2. Case Studies in Application: A series of computational experiments demonstrate the advantages of deep learning over traditional machine learning models across multiple areas, including operations management, inventory management, and risk management. For example, the use of deep learning in predicting sales volumes and insurance claims led to significant improvements in prediction accuracy and operational efficiency.
3. Advancement in Model Architecture: The authors propose a novel deep-embedded network architecture, designed to optimize the integration of categorical data within deep learning models. This architecture demonstrated superior performance in their case studies compared to both traditional models and standard DNN configurations.
4. Economic Implications: The paper provides a quantitative analysis of the economic benefits of employing deep learning techniques. Savings and enhanced decision-making capabilities were notable across the case paper applications, suggesting a substantial return on investment for businesses that embrace these technologies.

Practical and Theoretical Implications

The findings of this paper have multifold implications for the field of business analytics and operations research. Practically, they illustrate that companies stand to benefit considerably from adopting deep learning techniques, not only in terms of predictive accuracy but also through enhanced decision support systems that can lead to improved business performance and competitive advantage.

Theoretically, the paper suggests a need to develop more sophisticated models within business contexts that incorporate large datasets and complex feature interactions. The proposed deep-embedded architecture provides one such advancement, although the exploration of AutoML and transfer learning in this domain remains a promising area for future research.

Future Research Directions

The authors identify several avenues for future research. These include:

• Refinement of DNN architectures through automated machine learning (AutoML) methodologies to lessen the complexity of model tuning.
• Exploration of Bayesian deep learning models for uncertainty quantification.
• Development of interpretability techniques to enhance the transparency and trustworthiness of deep predictive models in a business context.

In conclusion, Kraus et al. make a compelling case for the expanded use of deep learning in business analytics and operations research. Their findings suggest that, while still in nascent stages, DNNs offer substantial advantages over traditional methods, warranting further research and encouraging adoption across various business sectors.