Perform Time-series Predictions in the R Development Environment by Combining Statistical-based Models with a Decomposition-based Approach
Zeydin Pala, Ahmet Faruk Pala · DSpace@MŞU · 2020
The analysis of a time-series (TS) measured or obtained by observing any area is an important step in characterizing a desired system or a phenomenon and predicting its future behavior. More precisely, predicting the value of an unknown variable is the objective of a predictive model used for TS. While doing this, it analyzes the relationships between past data well and reveals future predictions.In this study, the prediction method contrasts the decomposition-based approach with non-decomposition-based approaches. In the comparison process, prediction metrics for assessment, such as RMSE, MAE, MPE, and MAPE were used for method achievements and the results obtained were discussed.The experimental outcomes showed that the proposed decomposition-based approach performs better than non-decomposition-based approach in TS prediction processes.