基于人工免疫与支持向量机的日用水量预测  

Prediction of Daily Water Consumption Based on Artificial Immunity and Support Vector Machines

作　　者：岑健[1] 危阜胜[2] 

机构地区：[1]广东技术师范学院自动化学院,广州510635 [2]华南理工大学自动化科学与工程学院,广州510640

出　　处：《东莞理工学院学报》2009年第3期62-66,共5页Journal of Dongguan University of Technology

摘　　要：鉴于城市日用水量受众多因素的影响和具有复杂的非线性特点,提出一种基于人工免疫与回归支持向量机的用水量预测模型,根据免疫学原理的阳性选择对样本进行聚类后,由于去除非同类信息的干扰,样本数量大为减少,样本信息得到提纯.把模型应用于城市日用水量预测中,通过合理选择核函数、敏感系数、惩罚因子和宽度参数,从实例分析可知,该模型具有建模速度快、预测精度较高的特点.In view of the fact that the city daily water consumption is affected by many factors and has complicated non-linear characteristic, a water consumption prediction model is constructed based on support vector machines of regression analyses and artificial immunity. After using positive selection of immunology principle to cluster the samples, interference of non-same information is eliminated, with the number of samples greatly reduced and the information of samples purified. The model is applied to the prediction of city daily water consumption. By applying suitable kernel function, sensitivity coefficients, penalty factors and breadth parameters, the model is easily established and prediction precision gets improved.

关 键 词：支持向量机 用水量预测 回归模型 阳性选择 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]