Research on anomaly detection algorithm of time series data in cloud environment

Abstract

Cloud environment is a large-scale, distributed and complex system. Due to the complex interdependence and call relationship between its functional layers, the efficient operation and maintenance of cloud environment has become a major problem. The main form of daily monitoring KPI data in cloud environment is time series data. The prediction and anomaly detection of these data have always been two hot spots at home and abroad. The algorithm with high prediction accuracy and high anomaly detection accuracy can help us find the potential problems in the cloud environment and stop the loss in time to avoid large losses. Based on the previous relevant research results, this paper takes the data in the cloud environment as the research object, and takes improving the prediction accuracy and anomaly detection accuracy of the data as the research goal, and puts forward efficient and accurate prediction algorithms and anomaly detection algorithms suitable for the data characteristics in the cloud environment.