基于决策树算法的空气质量预测系统  

Air quality forecast system based on decision tree algorithm

作　　者：徐旭冉 涂娟娟[1] XU Xu-ran;TU Juan-juan(Jiangsu University of Science and Technology College,Zhenjiang 212003,China)

机构地区：[1]江苏科技大学计算机学院,江苏镇江212003

出　　处：《电子设计工程》2019年第9期39-42,共4页Electronic Design Engineering

摘　　要：针对目前空气质量预报多采用传统的数值模型现状,例如空气污染指数法,本次研究通过决策树算法以及大规模的训练数据集建立空气质量预测模型。传统的评估模型是在各种污染参数的污染分指数都计算出以后,取最大者为该区域或城市的空气污染指数固定数值区间的划分来评估空气质量。而基于决策树算法的空气质量评估模型通过采用自顶向下的递归方式对数据进行处理,把一个无序、无规则的实例集合归纳成一组树形结构表示的分类规则,得到了将所有污染参数作为评估空气质量因素的评估模型,可以有效的避免传统的空气质量预报模型的不灵活、边界值不准确的特点。同时可以根据季节和地区等外部因素构建不同的空气质量预测模型以应对外部因素的变化,从而可以构建完整,精确,现代化的空气质量智能预测系统。In view of the current situation of traditional numerical models,such as air pollution index method,this study establishes air quality prediction model through decision tree algorithm and largescale training data set. The traditional evaluation model is to evaluate the air quality by taking the largest as the fixed numerical interval of the air pollution index of the region or city after the pollution index of various pollution parameters has been calculated. And air quality assessment model based on decision tree algorithm by using the top-down recursive way to deal with data,put an unordered collection of instances,no rules induction into a set of tree structure according to the classification of the rules,get all the parameters for evaluating air quality pollution factor evaluation model,can effectively avoid the traditional air quality forecast model is not flexible,the characteristics of the boundary values are not accurate. At the same time,different air quality prediction models can be constructed according to external factors such as seasons and regions to cope with the changes of external factors,so as to build a complete,accurate and modernized air quality intelligent prediction system.

关 键 词：机器学习 决策树 空气质量预测 C4.5算法 信息增益 PYTHON 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]