Weather Forecasting Using Merged Long Short-term Memory Model

Afan Galih Salman, Yaya Heryadi, Edi Abdurahman, Wayan Suparta

Abstract


Over decades, weather forecasting has attracted researchers from worldwide communities due to itssignificant effect to global human life ranging from agriculture, air trafic control to public security. Although formal study on weather forecasting has been started since 19th century, research attention to weather forecasting tasks increased significantly after weather big data are widely available. This paper proposed merged-Long Short-term Memory for forecasting ground visibility at the airpot using timeseries of predictor variable combined with another variable as moderating variable. The proposed models were tested using weather timeseries data at Hang Nadim Airport, Batam. The experiment results showedthe best average accuracy for forecasting visibility using merged Long Short-term Memory model and temperature and dew point as a moderating variable was (88.6%); whilst, using basic Long Short-term Memory without moderating variablewasonly (83.8%) respectively (increased by 4.8%).

Keywords


Merged Long Short-term, Memory, Weather forecasting

Full Text: PDF

Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

Bulletin of EEI Stats