Abnormal index category identification method and device, storage medium and program product

The invention relates to the technical field of wireless, and particularly provides an abnormal index category identification method and device, a storage medium and a program product. The method comprises the following steps: acquiring time sequence data of a target key index, wherein the target key index is used for reflecting the network performance of a communication network cell; separating the time sequence data to obtain trend term data of the target key index, the trend term data being used for representing a time-varying rule of the target key index; performing anomaly detection on the trend term data to obtain abnormal data; and adopting a pre-trained classification model to predict the abnormal data, the time sequence data and the trend term data to obtain an abnormal category of the abnormal data. Compared with a diagnosis method based on a fixed threshold in the prior art, the flexibility and the accuracy are improved. 本公开涉及无线技术领域，尤其是提供一种异常指标的类别识别方法、设备、存储介质以及程序产品。该方法包括获取目标关键指标的时间序列数据，目标关键指标用于反映通信