基于网络搜索数据与随机森林模型的市场研究--以温州市商品住宅市场为例  

Market Research Based on Web Search Data and Random Forest Model—Take Wenzhou Commercial Housing Market as an Example

作　　者：张大力 林大川 郑一帆 ZHANG Dali;LIN Dachuan;ZHENG Yifan(Zhejiang college of Security Technology,Wenzhou 325024,China;College of Life Sciences,Zhejiang University,Hangzhou 310027,China)

机构地区：[1]浙江安防职业技术学院,浙江温州325024 [2]浙江大学生命科学学院,浙江杭州310027

出　　处：《现代信息科技》2022年第5期134-137,共4页Modern Information Technology

基　　金：温州市科研项目(R20210069)

摘　　要：以温州市房地产市场为研究对象,基于2011年1月至2020年12月间相关关键字的百度指数搜索数据和温州新建商品房销售价格,通过文本提取获取初始关键字,并二次拓展关键词库,后续使用相关时差分析筛选先行且具有相关性的关键字,再利用随机森林递归消除筛选出关键字,并根据筛选出的关键字建立随机森林模型,分析并预测,预测误差率在2.989%。This paper takes the real estate market in Wenzhou as the research object,based on the Baidu index search data of relevant keywords from January 2011 to December 2020 and the sales price of new commercial houses in Wenzhou,and obtains the initial keywords through text extraction,and expands the key thesaurus twice.It uses the relevant time difference to analyze and screen the first and relevant keywords,and uses the recursive elimination of random forest to screen the keywords,and establishes the random forest model according to the selected keywords for analysis and prediction.The prediction error rate is 2.989%.

关 键 词：网络爬虫 随机森林 温州房地产 网络搜索 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]