基于神经网络的磁平静期赤道电集流预测

利用BP神经网络技术分别对2008年后磁平静期印度扇区、秘鲁扇区以及CHAMP卫星的赤道电集流（EEJ）变化进行预测，其中神经网络训练数据为对应的2000—2007年磁平静期EEJ观测数据，输入参量为天数、地方时、太阳天顶角、太阳活动指数（F_10.7）、太阴时以及卫星地理经度，输出参量为EEJ.对EEJ预测结果进行了统计学分析，并且与实际观测结果进行对比.结果表明:BP神经网络对事件中EEJ的变化具有很好的预测能力，预测结果能够反映EEJ的重要分布特征；EEJ预测值与观测值之间具有很好的相关性，其中地磁台站观测值与预测值相关性系数可达85%以上.此外，将BP神经网络模型的预测结果与Yamazaki提出的经验模型结果进行对比，结果显示BP神经网络与其经验模型性能相当.研究结果表明，BP神经网络技术在平静期EEJ变化预测方面性能优异，具有良好的应用前景. 

Prediction of Equatorial Electro jet Based on the Neural Network during Quiet Time

In this study, BP neural network technology was used to predict the variations of Equatorial Electrojet (EEJ) of Indian sector, Peruvian sector and CHAMP satellite during the magnetic quiet pe-riod after 2008. The neural network training data is the corresponding observation data of EEJ during the period of magnetic quiet from 2000 to 2007. The input parameters are day of year, local time, solar zenith angle, solar activity index (F_10.7), lunar age and satellite geographic longitude, and the output parameters are EEJ. The EEJ prediction results are statistically analyzed and compared with the observation results. Results show that: BP neural network has good ability in predicting the variations of EEJ in the event, and the prediction results can reflect the important distribution characteristics of EEJ; there is very good correlation between the predicted values and the observed values of EEJ, and the correlation coefficient between the observed values and the predicted values of geomagnetic sta-tions can reach over 85%. In addition, the prediction results of BP neural network model are compared with those of Yamazaki's empirical model, and the performance of BP neural network is equivalent with Yamazaki's empirical model. The results show that BP neural network has excellent performance in predicting EEJ variations during quiet period, and has a good application prospect.