K. J. Albers, Morten Mørup, Mikkel N. Schmidt, Fumiko Kano Glückstad
{"title":"Predictive evaluation of human value segmentations","authors":"K. J. Albers, Morten Mørup, Mikkel N. Schmidt, Fumiko Kano Glückstad","doi":"10.1080/0022250X.2020.1811277","DOIUrl":null,"url":null,"abstract":"ABSTRACT Data-driven segmentation is an important tool for analyzing patterns of associations in social survey data; however, it remains a challenge to compare the quality of segmentations obtained by different methods. We present a statistical framework for quantifying the quality of segmentations of human values, by evaluating their ability to predict held-out data. By comparing clusterings of human values survey data from the forth round of European Social Study (ESS-4), we show that demographic markers such as age or country predict better than random, yet are outperformed by data-driven segmentation methods. We show that a Bayesian version of Latent Class Analysis (LCA) outperforms the standard maximum likelihood LCA in predictive performance and is more robust for different number of clusters.","PeriodicalId":50139,"journal":{"name":"Journal of Mathematical Sociology","volume":"46 1","pages":"28 - 55"},"PeriodicalIF":1.3000,"publicationDate":"2020-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1080/0022250X.2020.1811277","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Mathematical Sociology","FirstCategoryId":"90","ListUrlMain":"https://doi.org/10.1080/0022250X.2020.1811277","RegionNum":4,"RegionCategory":"社会学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"MATHEMATICS, INTERDISCIPLINARY APPLICATIONS","Score":null,"Total":0}
引用次数: 1
Abstract
ABSTRACT Data-driven segmentation is an important tool for analyzing patterns of associations in social survey data; however, it remains a challenge to compare the quality of segmentations obtained by different methods. We present a statistical framework for quantifying the quality of segmentations of human values, by evaluating their ability to predict held-out data. By comparing clusterings of human values survey data from the forth round of European Social Study (ESS-4), we show that demographic markers such as age or country predict better than random, yet are outperformed by data-driven segmentation methods. We show that a Bayesian version of Latent Class Analysis (LCA) outperforms the standard maximum likelihood LCA in predictive performance and is more robust for different number of clusters.
期刊介绍：
The goal of the Journal of Mathematical Sociology is to publish models and mathematical techniques that would likely be useful to professional sociologists. The Journal also welcomes papers of mutual interest to social scientists and other social and behavioral scientists, as well as papers by non-social scientists that may encourage fruitful connections between sociology and other disciplines. Reviews of new or developing areas of mathematics and mathematical modeling that may have significant applications in sociology will also be considered.
The Journal of Mathematical Sociology is published in association with the International Network for Social Network Analysis, the Japanese Association for Mathematical Sociology, the Mathematical Sociology Section of the American Sociological Association, and the Methodology Section of the American Sociological Association.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
