摘要
针对国内旅游人数预测研究了旅游人数的影响因素,讨论了输入层、隐含层、输出层等神经元的设置及网络训练的参数,综合考虑训练精度、训练时间、泛化能力等条件,动量—自适应学习速率调整算法是适合国内旅游人数预测的,并基于动量—自适应学习速率调整算法建立了神经网络模型;将模型应用于国内旅游人数预测系统,结果表明,该算法具有较好的准确性和鲁棒性,利用神经网络模型预测国内旅游人数是可行的.
In order to predict the number of domestic tourists,the related factors were studied,and the number of neurons in input layer,output layer,hidden layer and network training parameter settings were discussed;the training time,training accuracy,generalization ability and other conditions were considered.The algorithm of momentum-adaptive learning rate adjustment the most appropriate to predict domestic tourists.The model is applied to forecasting system.Its actual operation shows that the algorithm is accurate and robust,and the application of neural network models to the prediction of domestic tourists is feasible and effective.
出处
《甘肃科学学报》
2012年第1期81-83,共3页
Journal of Gansu Sciences
基金
国家自然科学基金项目(0211003026/11220300)
国家重点基础研究发展规划(973)项目(041J007026/21010703)
关键词
神经网络
国内旅游人数
预测
neural network
the number of domestic tourists
prediction
