A hybrid ABC-SVM approach for multi-dimensional data classification with synthetic data balancing

User behaviour data plays a vital role in digital decision-making, especially in education, finance, and healthcare. However, traditional methods often fail to capture the complex characteristics of user behaviour, perform poorly on multi-dimensional data, and struggle with class imbalance, which limits model performance. To overcome these challenges, this study constructs a dynamic user behaviour dataset from the Chaoxing system and adopts the synthetic minority oversampling technique (SMOTE) to address data imbalance problem. The artificial bee colony (ABC) algorithm is combined with the support vector machine (SVM) to optimise model parameters and improve performance. Experimental results show that the proposed ABC-SVM model performs well in complex classification tasks with an accuracy of 97.9%, outperforming baseline and other optimisation methods. This study highlights the potential of intelligent optimisation algorithms in multi-dimensional data analysis and provides a reference for intelligent systems in other fields.