摘要
针对房价波动大的数据特征,将多小波分析与房价预测问题结合,以北京市2010-2018年的房屋数据作为研究对象,探究了利用Haar小波变换、Daubechies系列小波变换以及基于过采样预处理的GHM多小波变换和CL多小波变换处理房价数据的分解重构效果,并通过对高频系数进行门限阈值量化重构处理以达到去噪的目的;建立支持向量机(SVM)预测模型,通过探究4种小波处理方法对房屋价格预测的影响结果,给出了相应预测效果更佳的数据处理方法和选择依据。
In recent years,housing problems have become more prominent.However,due to the large fluctuations in data,it is difficult to predict the price of the housing market.By focusing on the data characteristics of house prices,this paper combines multi-wavelet analysis with house price forecasting.Using Beijing house price data,the effects of manipulating these data with Haar,Daubechies wavelet transform and GHM,and a CL multi-wavelet transform based on oversampling was compared.By means of a support vector machine ( SVM),a model is established to predict unit-price based on treating the data with different wavelet transforms.Finally,a strategy for selecting house price data manipulation methods is suggested.
作者
邬嘉怡
王思玉
史宏炜
李虎森
楼凯达
崔丽鸿
WU JiaYi;WANG SiYu;SHI HongWei;LI HuSen;LOU KaiDa;CUI LiHong(Faculty of Science,Beijing University of Chemical Technology,Beijing 100029,China)
出处
《北京化工大学学报(自然科学版)》
CAS
CSCD
北大核心
2019年第5期101-106,共6页
Journal of Beijing University of Chemical Technology(Natural Science Edition)
基金
国家级大学生创新训练计划(201810010050)