Time series forecasting using exponential smoothing to predict the number of website visitor of Sebelas Maret University

Website is one of the technology services which used by the Sebelas Maret University (UNS) as information media. Related with IT infrastructure that should be prepared, the management of UNS requires the estimated number of website visitors in the future. This estimated number of website visitors can be predicted through forecasting technique. The objective of this study is to predict the UNS website visitor using exponential smoothing. Exponential smoothing is chosen because quite simple and also has advantages in being able to adapt the change trends and seasonal components. The optimum smoothing constants obtained by calculation of 0.01 interval increments on 60 training data in 2009 until 2013, the smallest error value obtained is 12.95% with smoothing constants alpha = 0.24, beta = 0.10 and gamma = 0.01.The results of the evaluation using 6 tester data in January 2014 until June 2014 shows this forecasting model is categorized as good forecasting with an error calculation (MAPE) value as 10.27% Based on forecasting results, predicted that the number of website visitors of UNS will increase by 13.62% per year, which in July would be the highest seasonal index with average index value as 1.53 and November would be the lowest seasonal index with average index value as 0.64.