Abstract:
This study introduces a novel approach to predicting the success of small businesses in Kazakhstan, leveraging Graph Neural
Networks (GNNs) to analyze a comprehensive set of business parameters. Recognizing the critical role of small businesses
in the national economy, this research aims to provide stakeholders with a predictive tool that utilizes advanced machine
learning techniques to evaluate business outcomes. By integrating data on revenue, number of employees, market dynamics,
and other key operational metrics, the model captures the complex interactions within the business ecosystem. The
methodology involves constructing a graph-based representation of the business landscape, where nodes represent individual
businesses and edges denote relationships and influences among them. Through this framework, the GNN model learns to
identify patterns and predictors of success, offering insights that traditional linear models might overlook. Preliminary results
indicate a strong correlation between specific business parameters and their likelihood of success, highlighting the potential
of GNNs in strategic decision-making. This paper not only contributes to the academic discourse on predictive analytics in
business but also proposes a practical tool for entrepreneurs, investors, and policymakers in Kazakhstan to foster a thriving
small business sector. Future work will focus on refining the model, incorporating real-time data, and expanding its
applicability to other regions and sectors.
