Application and anomaly detection of predictive model based on time series
XU Yinlong · Computer-aided engineering · 2006
Due to the problem that errors are often found in prediction data generated by the method of time series because of the uncertainties when it takes place, a method of anomaly detection based on iteration is developed by analyzing the historic time series data. It can find the possible abnormal events and erase the errors in them. The tests prove that it can greatly improve the prediction accuracy, and the prediction data can provide more reference to the future works.