基于支持向量机的城市空气质量时间序列预测模型探究

摘要

Air pollution is a serious problem in the moment. Carrying out the air quality monitor and fore-cast has a great significance to pollution control and harm reduction.Support vector machine model is a good instrument which can be using to forecast regression as well as time series.This paper takes radial basic function as a kernel function, and uses cross validation to optimize the parameter thus to support vector machine time series model.By using one city’s air quality index from January to August in 2013 as air quality parameter to do the practical analysis,the author found this model had a quite good predictive effect,which had a certain practical value.%空气污染问题在当下是一个十分严重的问题。开展空气质量监测、预测工作对于污染控制，降低危害具有重要意义。支持向量机模型是进行回归预测性能良好的工具，并可用于时间序列预测。文章采用径向基函数作为核函数，用交叉验证的方法优化参数构造支持向量机时间序列预测模型，选取某地市2013年1月至8月的空气质量指数作为空气质量参数进行实证分析，表明模型预测效果很好，具有一定实用价值。