基于时间序列延迟相关算法改进LSTM的臭氧浓度预测模型

Improved LSTM-Based Ozone Concentration Prediction Model Based on Time Series Delay Correlation Algorithm

摘要: 臭氧污染日益引起人们的重视，如何准确预报臭氧浓度成为一个重要课题。以杭州市多个站点臭氧的小时监测数据为研究对象，分析不同站点的臭氧浓度变化的相关性，结合长短期记忆(LSTM)神经网络模型提出一个基于时间序列延迟相关算法改进LSTM的臭氧浓度预测模型，通过与传统的LSTM模型和SpaceLSTM模型进行对比实验，结果表明：所提出的方法均方误差最小，预测结果更加准确。

Abstract: Ozone pollution has attracted increasing attention. How to accurately predict ozone concentration has become an important subject. Taking the hourly monitoring data of ozone at multiple sites in Hangzhou as the research object, the correlation of ozone concentration changes at different sites was analyzed, and a long-term short-term memory (LSTM) neural network model was proposed to improve the LSTM ozone concentration prediction based on a time series delay correlation algo-rithm. The model is compared with the traditional LSTM model and SpaceLSTM model, and the results show that the proposed method has the smallest mean square error and the prediction result is more accurate.