Optimization and Automation of Multi-Layered EnsembleLearning for Short-Term Forecasting in Agro-Climatology
Résumé
Agriculture is one of the areas whose activities heavily depend onweather forecasts. This paper, using Spark, proposes OptMLEL, an improvedversion of our previous work (i.e., MLEL) for short-term forecasting to assistfarmers in their decision-making. OptMLEL helps to resolve the computationalbottleneck when training MLEL in a single computational node. It also ap-plies new features in training, tunes the algorithm parameters automatically incase of adding/removing data source providers and selects the number of lay-ers automatically based on the amount of data. The obtained results show theapplicability and performance of the proposed method.
Domaines
| Licence |
|---|