الفهرس 
A low footprint weather forecasting framework using a single-layered seasonal attention encoder-decoder model
SummaryOnly 14 pages are availabe for public view
 |  | from | 164 |  |  |
| | | from | 164 | | |
Abstract
This thesis presents a lightweight deep learning model for weather forecasting. The model is an encoder-decoder with a seasonal attention mechanism. This model is enclosed in a framework for training the model and testing it, and deploying it onto a low-cost microcontroller for use in the remotely located olive groves. Several experiments are conducted on the model to test its predictive performance power. A prototype is built for use in the real-life weather forecasting scenario and is tested.