将调查抽样理论应用于网络抓取数据：使用 IPW 估计量分析 OBEC 数据.pdf

《将调查抽样理论应用于网络抓取数据：使用 IPW 估计量分析 OBEC 数据.pdf》由会员分享，可在线阅读，更多相关《将调查抽样理论应用于网络抓取数据：使用 IPW 估计量分析 OBEC 数据.pdf（22页珍藏版）》请在三个皮匠报告上搜索。

1、DraftWIN Conference.From Web to Data.3-4 February 2025,Gdask,PolandApplying Survey Sampling Theory toWeb-Scraped Data:An Analysis ofOBEC Data Using the IPW Estima-torVilma Nekraait-LiegDraftOutlineIntroductionBackgroundResearch goal and problemsThe use of survey sampling theoryData availability scen

2、ariosIPW estimatorPractical Implementation3-4 February 2025,Gdask,PolandThe use of IPW estimator for OBEC datap.2 of 22DraftOutlineIntroductionBackgroundResearch goal and problemsThe use of survey sampling theoryData availability scenariosIPW estimatorPractical Implementation3-4 February 2025,Gdask,

3、PolandThe use of IPW estimator for OBEC datap.3 of 22DraftBackground(ICT survey)The aim of the European ICT usage surveys is to collect and disseminateharmonised and comparable information on the use of Information andCommunication Technologies in enterprises and e-commerce at Europeanlevel.Frequenc

4、y of data collection:annual.Population:enterprises with 10 or more persons employed.Statistical unit:enterprise.Breakdown:by size class,by NACE Rev 2 categories.3-4 February 2025,Gdask,PolandThe use of IPW estimator for OBEC datap.4 of 22DraftURL search algorithm(1)*URL list in 2021 was provided by

5、private company,for the other year the previousyear list is used.3-4 February 2025,Gdask,PolandThe use of IPW estimator for OBEC datap.5 of 22DraftURL search algorithm(2)Web scrape1.Selenium module is used in Python.2.All data are saved in the sqLite database.3.More than 10 different search phrases(

6、Enterprise ID,name,contactinformation)are used to check if this is the right page.URL finder1.Sending search terms to a search engine.2.Scraping the result URLs.3.Extracting the scraped data.4.Creating a machine learning or rule-based model to link websites toenterprises:Logistic regression,random f