Towards a Framework for Intelligent Sampling: Comprehensive Review of Challenges, AI Techniques, and Tools

Intelligent data sampling is an innovative method that enhances conventional data sampling procedures by utilizing machine learning and artificial intelligence approaches. In this manuscript, we deep dive into the scientific and grey literature to find the main challenges faced by data sampling and elaborate on the various AI techniques utilized to mitigate them. We identify key issues such as class imbalance, overfitting, computational inefficiency, and bias, which often hinder traditional sampling methods. Furthermore, we explore AI-driven techniques that have been integrated into the sampling process to address these challenges effectively. As a result, we propose a novel framework for intelligent sampling that incorporates an AI-powered recommender system. This system dynamically selects the most appropriate sampling technique based on the specific characteristics of the data and the needs of the predictive model. By automating and optimizing the selection of sampling methods, our framework aims to enhance model performance, improve resource efficiency, and adapt to diverse real-world applications.