长短期记忆神经网络预测中国智能照明市场规模

Prediction of the Scale of China's Intelligent Lighting Market Based on LSTM

由于我国智能照明市场规模数据序列呈非凸凹一致、非光滑分布的特点,传统循环神经网络预测效果差强人意,因此,采用性能更优的长短期记忆神经网络模型(LSTM)对其进行预测。结果显示:LSTM神经网络的平均预测误差仅为0.2760%,比BP神经网络的0.7841%减小了64.8004%,比RNN神经网络的0.5611%减小了50.8109%。运用LSTM神经网络对我国2022-2025年智能照明市场规模进行了预测。通过分析表明,这一预测结果有较高的可信度。

Because the data series of scale of Chinas intelligent lighting market has the characteristics of non convex concave consistent and non smooth distribution,the predicted effect of traditional neural network is not satisfactory.Therefore,long-term and short-term memory neural network is used to predict it.The results show that the average prediction error of the LSTM neural network is only O.2760%,which is 64.8004%less than the 0.7841%of BP neural network and 50.8109%less than the 0.5611%of the RNN neural network.The LSTM neural network was used to predict the scale of Chinas intelligent lighting market from 2022 to 2025.Through analysis,it is shown that this prediction result has high reliability.