Call Forecasting Based on SARIMA and SVM Hybrid Model

Call Forecasting is the premise of staffing and scheduling in call center. This paper is based on the analysis of actual data and the comparison of various time series forecasting methods , proposed the hybrid algorithm which combining the SARIMA model and support vector machine model. We used the SARIMA(seasonal autoregressive integrated moving average) model with 48 periods and a input for the linear part of time series. Taking into account the deficiencies that the statistical prediction algorithm as a linear data model can not capture nonlinear data, we used the Support Vector Machine model to fit the residuals of SARIMA to complement the predictive value of the nonlinear part, which leads to better analysis and prediction results.