An Optimization of Data Sourcing Services using Data Gateway

Abstract

Introduction: This article presents the design and testing of a GNN-based data gateway that can be used to build query-based data mapping and categorisation in corporate settings.  By using the Graph Neural Network's (GNN) capacity to record complex interactions between entities inside datasets, the proposed method aims to increase data retrieval speed and accuracy.

Objectives: The study aims to improve commercial applications' decision-making processes by optimising data sourcing services using a GNN-based data gateway, boosting query-based data mapping's accuracy and efficiency.

Methods: Five data classes are used to evaluate the GNN model: MongoDB, Sales, Product, Employee, and Customer.  Its performance is evaluated using important measures including F1-score, accuracy, precision, and recall in comparison to more conventional machine learning models, especially Random Forest (RF) and Logistic Regression (LR).

Results: Across all assessment measures, the GNN outperformed the LR and RF models.  With a total accuracy of 0.99, it demonstrated exceptional recall and precision.  The reason for its exceptional performance is its capacity to record complex relationships between data nodes.

Conclusions: These findings demonstrate the GNN-based data gateway's scalability and dependability as a potential strategy for contemporary data-driven commercial applications.  In a variety of sectors, it has significant potential for improving query-based data mapping, facilitating improved decision-making, and improving operational efficiency