Abstract | ||
---|---|---|
Genetic programming (GP) is an evolutionary algorithm that has received a lot of attention lately due to its success in solving hard world problems. There has been a lot of interest in using GP to tackle forecasting problems. Unfortunately, it is not clear whether GP can outperform traditional forecasting techniques such as auto-regressive models. In this contribution, we present a comparison between standard GP systems qand auto-regressive integrated moving average model and exponential smoothing. This comparison points out particular configurations of GP that are competitive against these forecasting techniques. In addition to this, we propose a novel technique to select a forecaster from a collection of predictions made by different GP systems. The result shows that this selection scheme is competitive with traditional forecasting techniques, and, in a number of cases it is statistically better. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1007/s11047-015-9536-z | Natural Computing: an international journal |
Keywords | Field | DocType |
Genetic programming,Time series forecasting,Auto-regressive models,M1 and M3 competitions | Exponential smoothing,Time series,Evolutionary algorithm,Genetic programming,Artificial intelligence,Moving-average model,Machine learning,Mathematics | Journal |
Volume | Issue | ISSN |
16 | 1 | 1567-7818 |
Citations | PageRank | References |
4 | 0.54 | 14 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Mario Graff | 1 | 125 | 21.24 |
Hugo Jair Escalante | 2 | 939 | 73.89 |
Fernando Ornelas-Tellez | 3 | 76 | 13.03 |
Eric Sadit Tellez | 4 | 61 | 15.04 |