Developing the Hybrid Forecasting Model on the Short-Term Time Series

Van-Chung Pham, Ngoc-Canh Pham
2020 ICIC Express Letters  
The problem of time series data forecast has many practical benefits, thus getting more and more attention from many people. Doing forecast on short-term time series is a practical problem today. There are two commonly used methods: the exponential smoothing and artificial neuron network (ANN). However, many studies show that for trend and seasonal data, the quality of the ANN is not high. Similarly, for non-linear data, the quality of the exponential smoothing is not good. In this paper, we
more » ... posed a new model to forecast short-term time series data by developing a hybrid model between the ANN and exponential smoothing. Empirically, this hybrid model can forecast on different data types and give good forecast results.
doi:10.24507/icicel.14.10.1017 fatcat:sywaazclhvgrjfeb4zbvosfk7u