Content area
This paper investigates the impact of data preprocessing on the performance, efficiency, and environmental footprint of AI models in cloud-based applications, focusing on a case study involving healthcare applications such as chronic disease detection. We analyze how preprocessing techniques affect some of the most commonly used Machine Learning (ML) algorithms, namely K-means, SVM, and KNN, emphasizing their role in reducing computational load, energy consumption, and carbon emissions in data centers. Our results demonstrate that the impact of preprocessing on both accuracy and processing speed varies depending on the algorithm and the type of preprocessing applied. Notable improvements in precision and processing time reductions of up to 35% were observed, highlighting the potential of preprocessing to enhance the performance and sustainability of ML algorithms.