Five-axis CNC machine tools are indispensable for manufacturing parts with complex geometry, yet the conversion of raw kinematic measurements into reliable maintenance knowledge is still often fragmented and manual. This article presents a Python-based monitoring environment that reads R-Test export files, invokes one of three selectable machine-learning models – Multilayer Perceptron (MLP), Kolmogorov-Arnold Network (KAN), or Multi-Output Gaussian Process (MOGP) – and automatically produces bilingual diagnostic PDF reports. A heuristic rule layer interprets displacement trajectories and assigns probable fault categories based on signal signatures, without relying on labelled fault datasets. Validation on 1568 R-Test records from two five-axis milling machines shows that MOGP remains the most accurate inference backend. The main contribution of the paper is the integration of data ingestion, inference, heuristic interpretation, visual analytics, and report generation into a deployable condition-monitoring platform rather than a standalone model-benchmarking study.
Science Journal 