Time-series prediction of hourly atmospheric pressure using ANFIS and LSTM approaches.

ATMOSPHERIC pressure; ATMOSPHERIC sciences; ENVIRONMENTAL sciences; MACHINE learning; CLIMATOLOGY

Atmospheric pressure (AP), which is an indicator of weather events, plays an important role in climatology, agriculture, meteorology, atmospheric and environmental science, human and animal life, and Earth's living ecosystem. In this regard, accurate AP forecasting plays a crucial role in today's life as it provides critical information about future weather events. In this study, four different machine learning techniques such as long short-term memory (LSTM) neural network, adaptive neuro-fuzzy inference system (ANFIS) with fuzzy c-means, ANFIS with subtractive clustering, and ANFIS with grid partition (GP) were used for one-hour-ahead AP forecasting. To achieve this, the hourly AP data measured between 2012 and 2019 at the seven measurement stations (Adana, Ankara, Gumushane, Denizli, Kirklareli, Sanliurfa, and Van) in different climate regions of Turkey were obtained. The estimation accuracy was verified by four performance criteria: R, RMSE, MAPE, and MAE. As a result, the highest relative R-value of 0.9986 and the lowest error values of RMSE = 0.2905 hPa, MAPE = 0.0230%, and MAE = 0.2040 hPa for one-hour-ahead AP forecasting were obtained from the ANFIS-GP model. [ABSTRACT FROM AUTHOR]

Copyright of Neural Computing & Applications is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)