Effective prediction of web services response time has important guiding significance for service providers to guarantee the quality of service, using big data approach to process a large number of indictors’ historical data can improve the efficiency of prediction. Correlation analysis is proposed to remove indicators’ items that are not highly correlated with response time, the computed data volume is reduced by the feature dimension reduction, and the multi layer LSTM optimization algorithm with dynamic adjustment parameters is designed to predict the response time of web services. Experiments show that the proposed method can predict the response time of Web services efficiently and accurately.
