This paper discusses the increasing amount of data handled by companies and the need to use Big Data and Data Analytics to extract value from this data. However, due to the large amount of data collected, challenges related to the computational capacity of machines often arise when performing this analysis to acquire relevant information for the organization, especially when we are using edge computing. The paper aims to train machine learning models using compressed data, with two compression techniques applied to the original data. The results show that models trained with compressed data achieved similar accuracy to those trained with uncompressed data, and different compression techniques were compared. The research extended a previous study by analyzing the use of autoencoders for compression and reducing both instances and dimensionality of the dataset. The accuracy rate of the models when trained with compressed data instead of original data was maintained.

Fractional-order systems capture complex dynamic behaviors more accurately than integer-order models, yet their real-time identification remains challenging, particularly in resource-constrained environments. This work proposes a hybrid framework that combines Particle Swarm Optimization (PSO) with various artificial intelligence (AI) techniques to estimate reduced-order models of fractional systems. First, PSO optimizes model parameters by minimizing the discrepancy between the high-order system response and the reduced model output. These optimized parameters then serve as training data for several AI-based algorithms—including neural networks, support vector regression (SVR), and extreme gradient boosting (XGBoost)—to evaluate their inference speed and accuracy. Experimental validation on a custom-built heating system demonstrates that both PSO and the AI techniques yield precise reduced-order models. While PSO achieves slightly lower error metrics, its iterative nature leads to higher and more variable computation times compared to the deterministic and rapid inference of AI approaches. These findings highlight a trade-off between estimation accuracy and computational efficiency, providing a robust solution for real-time fractional-order system identification on edge devices.

The industrial sector has undergone significant digital transformation, driven by advancements in technology and the Internet of Things (IoT). These developments have facilitated the collection of vast quantities of data, which, in turn, pose significant challenges for real-time data processing. This study seeks to validate the efficacy and accuracy of edge computing models designed to represent subprocesses within industrial environments and to compare their performance with that of traditional cloud computing models. By processing data locally at the point of collection, edge computing models provide substantial benefits in minimizing latency and enhancing processing efficiency, which are crucial for real-time decision-making in industrial operations. This research demonstrates that models derived from distinct subprocesses yield superior accuracy compared to comprehensive models encompassing multiple subprocesses. The findings indicate that an increase in data volume does not necessarily translate to improved model performance, particularly in datasets that capture data from production processes, as combining independent process data can introduce extraneous ‘noise’. By subdividing datasets into smaller, specialized edge models, this study offers a viable approach to mitigating the latency challenges inherent in cloud computing, thereby enhancing real-time data processing capabilities, scalability, and adaptability for modern industrial applications.