基于LSTM神经网络的大连市空气质量指数预测

Prediction of Air Quality Index in Dalian Based on LSTM Neural Network

摘要: 由于大气环境的复杂性和多变性，用传统方法预测空气质量指数精度较低，本文首先对空气质量指数和主要污染物进行相关性分析，再采用基于TensorFlow的长短期记忆神经网络(LSTM)模型，对大连市的空气质量指数进行预测，并进行误差分析。实验结果表明，与支持向量机模型和BP神经网络模型相比，LSTM神经网络模型预测空气质量指数的精度较高。

Abstract: Due to the complexity and dynamics of atmospheric environment, the accuracy of air quality index prediction by traditional method is low. Firstly, this paper analyzes the correlation between air quality index and main pollutants. Then, the long and short-term memory neural network model based on Tensor Flow is used to predict the air quality index of Dalian and the error analysis id made. The experimental results show that LSTM neural network model is more accurate than support vector machine model and BP neural network model in predicting air quality index.