Use of ANN models in the prediction of meteorological data
Use of ANN models in the prediction of meteorological data
复制标题
ANN 模型在气象数据预测中的应用
DOI：
10.1007/s40808-019-00590-2
复制
发表时间：
2019
影响因子：
3
通讯作者：
R. Boadh
中科院分区：
文献类型：
--
作者：
P. Rajendra;K. Murthy;A. Subbarao;R. Boadh
本研究是利用人工神经网络（ANN）模式的气象资料的预测。人工神经网络模型，即多层感知器（MLP）和径向基函数（RBF）已被用于气象数据的预测。为了确认模型的性能，每小时和每月的预测变量进行了比较，通过多元线性回归模型，气象站记录的结果。MLP和RBF在所有情况下的预测准确率为91-96%。此外，预测显示出与记录的数据在0.61和0.94之间的强线性相关性。为人工神经网络作为气象预报的有力工具提供了保证。调查是作为一个案例研究，位于印度的两个气象站。本研究的一个扩展是将这些人工神经网络应用于其他地区的不同数据类型的气象数据将是未来工作的兴趣。
The present study is to use artificial neural network (ANN) models for the prediction of meteorological data. Artificial neural network models namely multiple layer perceptron (MLP) and radial base function (RBF) have been used for the prediction of meteorological data. To confirm the performance of the models, the hourly and monthly predictions of variables have been compared with results obtained by multi-linear regression model, recorded by meteorological stations. The MLP and RBF have given 91–96% accuracy for predictions of all cases. In addition, the forecasts demonstrated a strong linear correlation with the data recorded in between 0.61 and 0.94. The present work has given assurance to use artificial neural network as a strong tool to predict the meteorological data. The investigation is conducted as a case study of two meteorological stations situated in India. An extension of the present study is to apply these ANNs in other regions with different data types of meteorological data will be the interest of future work.