基于时间序列嵌入的混合特征学习用于设备异常预测

In predictive maintenance of equipment, deep learning-based time series anomaly detection has gained significant attention; however, pure deep learning methods often struggle to achieve adequate accuracy on real-world data. This study proposes a hybrid approach that combines 64-dimensional time series embeddings from the Granite TinyTimeMixer with 28-dimensional statistical features derived from domain knowledge for HVAC equipment anomaly prediction tasks. Specifically, time series embeddings extracted from a Granite TinyTimeMixer encoder fine-tuned with LoRA (Low-Rank Adaptation) are integrated with 28 types of statistical features, including trend, volatility, and drawdown indicators, which are then learned using a LightGBM gradient boosting classifier.