The evolution of mobile communication networks towards the era of 5G necessitates robust anomaly detection systems to ensure uninterrupted user experience and network efficiency. This paper introduces a novel approach using Deep Generalized Canonical Correlation Analysis (DGCCA) for anomaly detection in User Equipment (UE) parameters. DGCCA, an extension of Canonical Correlation Analysis, offers the ability to capture complex correlations among diverse UE parameters, surpassing the limitations of conventional anomaly detection methods. Leveraging a meticulously curated simulated dataset derived from an urban area scenario, this study evaluates DGCCA alongside traditional models such as SVM, Naive Bayes, Random Forests and FCNN. Results showcase DGCCA’s remarkable accuracy of 92.23% with a low false negative rate of 1.55%, surpassing its counterparts. The outcomes underscore DGCCA’s robustness in fortifying self-organizing networks, offering a promising solution for resilient and adaptive mobile communication systems in the modern era. Incorporating this anomaly detection system into the ORAN framework has the potential to improve the forecasting of optimized base station parameters.