Multi-step LSTM Prediction Model for Visibility Prediction
| Autor: | Yunlong Meng, Yao Xiao, Xian Yuan, Heng Zuo, Bo Chen, Fengliang Qi |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Sequence
010504 meteorology & atmospheric sciences Computer science business.industry Deep learning Visibility (geometry) 010502 geochemistry & geophysics computer.software_genre 01 natural sciences Data set Artificial intelligence Data mining Time series business Focus (optics) computer 0105 earth and related environmental sciences |
| Zdroj: | IJCNN |
| DOI: | 10.1109/ijcnn48605.2020.9206744 |
| Popis: | In this paper, we present a deep learning framework with attention mechanism for visibility prediction. We firstly formulate visibility prediction as a temporal prediction problem. An encoder-decoder architecture based network is proposed to generate a multi-step prediction. To adaptively focus on different parts of the input and output sequence, we incorporate input attention and temporal attention into the network. Experiments verify the feasibility of the proposed model. We produce state-ofthe-art prediction accuracy (68.9%) on the runway visual range prediction in our customized data set collected at observation stations of the airport. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|