基于LSTM神经网络模型的钢铁价格预测
【摘 要】由于钢铁价格具有影响因子难以确定和非线性的特点，在数据挖掘预测分析是，传统的预测方法只能够对钢铁价格进行小数据量的分析，导致预测精度低、速度慢且效率低下。随着大数据的深入研究，将神经网络与spark相结合，能满足用户对实时数据处理的需求。在多个深度学习神经网络模型中，基于长短期记忆单元（Long Short-term memory，LSTM）的递归神经网络（recurrent neural network，RNN）模型因为其能有效利用序列数据中长距离依赖信息的能力，非常適用于价格指数的预测中。文章利用python和lstm，结合近几年钢铁交易价格的走势数据，对数据进行回归拟合，生成训练模型，然后将得出的模型用来对未来的钢铁交易价格进行预测，使用均方误差（MSE）对预测数据和原始数据进行误差分析与处理，并与支持向量回归（SVR）模型进行对比。
【关键词】大数据；LSTM；RNN；MSE；SVR；价格指数预测
【Abstract】Because of the characteristics of steel prices is difficult to determine the impact factor and nonlinear， in data mining prediction analysis， analysis of the traditional prediction methods can only be a small amount of data on steel prices， resulting in low prediction accuracy， slow speed and low efficiency.With the deep research of large data， the combination of neural network and spark can meet the needs of users for real-time data processing. In multiple depth learning neural network models， The recurrent neural network model based on long and short memory units is very suitable for the prediction of price index because it can effectively utilize the ability of long distance dependence information in sequence data. The use of Python and LSTM， combined with the trend of steel price data in recent years， the data fitting， the generation of the training model， then the model is used to predict the future price of steel trading， analysis and handle the error of forecast data and the original data using mean square error， and wi...
|
== 试读已结束，如需继续阅读敬请充值会员 ==
|
|
本站文章均为原创投稿，仅供下载参考，付费用户可查看完整且有格式内容！
(费用标准：38元/2月，98元/2年，微信支付秒开通！) |
| 升级为会员即可查阅全文 。如需要查阅全文，请 免费注册 或 登录会员 |
|
