Predictive modeling in supply chain management based on artificial intelligence methods

Aim.

The work aimed to identify and compare comprehensively predictive modeling methods using artificial intelligence in supply chain management based on a systematic literature review.

Objectives.

The work seeks to analyze the main predictive modeling methods; to conduct a systematic literature review, including defining criteria for selecting studies being analyzed, defining the study design, and selecting a sample in a step-wise manner; to compare the advantages and disadvantages of artificial intelligence and traditional statistical methods for predictive modeling in supply chain management; to compile recommendations for the implementation and application of various predictive modeling methods depending on the type of supply chain operations.

Methods.

The study is based on systematic literature review (SLR). The author applied the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standard which ensures transparency, structuredness, and methodological rigor in the review process. The study methodological framework consists of several components which include formulating research questions, conducting a comprehensive search of relevant sources, applying developed criteria for selecting materials, and integrating the obtained data to form insightful conclusions.

Results.

The systematic analysis results reveal that the use of machine learning methods is becoming increasingly widespread and has been proven to offer broad perspectives for improving decision-making and forecasting accuracy in SCM. The study offers recommendations for the use of predictive modeling methods based on supply chain operations. In addition to the benefits of using artificial intelligence in SCM, several shortcomings have been identified, particularly issues related to data quality, model interpretability, and the need for domain knowledge. Finally, a summary of the results shows that while AI-based predictive models can improve efficiency and responsiveness in supply chain management, their successful implementation requires careful consideration of organizational context and operational constraints.

Conclusions.

A hybrid approach to predictive supply chain analytics is currently the most applicable. This approach combines traditional statistical methods with machine learning techniques, as it enables multi-stage data validation and processing, mitigating issues of interpretability and quality.