This invention discloses a data-augmented anomaly localization method for IoT-sensing
cloud data centers, relating to the field of
cloud data. The method includes: S1 acquiring a training dataset; S2 constructing an
anomaly detection model; S3 importing the training dataset into the
anomaly detection model for optimization training; S4 acquiring the
raw data from the
cloud data center in real time and importing it into the optimized
anomaly detection model to obtain predicted values for the
raw data; and S5 determining anomalies in the
raw data of the
cloud data center based on the predicted values and the true values of the raw data. The method utilizes data augmentation-neural transformation and convolutional long short-
term memory networks to mine different aspects of the temporal and spatial characteristics of multivariate
time series, increasing the amount of training data and reducing false positives caused by non-stationary and nonlinear real-time IoT data. Furthermore, it employs an attention-based autoregressive LSTM network to reduce the dimensionality of the data and extract features again, fully acquiring information beneficial to prediction from the data and improving the robustness of the model's anomaly detection.