当前位置: X-MOL 学术Cognit. Comput. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
The Big-2/ROSe Model of Online Personality
Cognitive Computation ( IF 5.4 ) Pub Date : 2021-07-29 , DOI: 10.1007/s12559-021-09866-1
Gerardo I. Simari 1, 2, 3 , Fabio R. Gallo 1, 2 , Marcelo A. Falappa 1, 2 , Maria Vanina Martinez 4, 5
Affiliation  

The Big-5/OCEAN personality traits model, one of the central approaches to psychometrics, has been shown to have many applications over a variety of disciplines. In particular, correlations have been studied leading to effective characterization of people’s behavior, and the model has become notorious for its role in the Cambridge Analytica/Facebook scandal surrounding the 2016 US presidential elections. In this paper, we develop Big-2 (or ROSe, for Relationship to Others and to Self), a model via which the personality of users of online platforms can be studied using a lightweight set of markers focused on online behavior, avoiding the major data privacy pitfalls afflicting approaches based on more powerful models that characterize personal aspects of the human psyche. Evaluation of Big-2’s effectiveness is done in two parts: a quantitative evaluation on a specific prediction task and a qualitative one based on an analysis of the different ways in which the Big-2 traits can be derived from online behavior, proposing a general template to guide such efforts. Quantitative results show that our lightweight model can match or surpass the performance of Big-5 in a prediction task, while qualitative results show that it is feasible to implement the model based on the observation of basic online user behavior. Our main result is a general-purpose model that can be used to characterize the personality traits of users of online platforms in an ethical manner. Our proposed model provides a valuable tool to carry out effective and explainable analyses of online personality, avoiding the collection of unnecessary user data that would open the possibility for ethical violations.



中文翻译:

在线个性的 Big-2/ROSe 模型

Big-5/OCEAN 人格特质模型是心理测量学的核心方法之一,已被证明在各种学科中有许多应用。特别是,研究了相关性以有效表征人们的行为,并且该模型因其在围绕 2016 年美国总统选举的 Cambridge Analytica/Facebook 丑闻中的作用而臭名昭著。在本文中,我们开发了 Big-2(或 ROSe,表示与他人和自我的关系),该模型可以使用一组专注于在线行为的轻量级标记来研究在线平台用户的个性,避免主要数据隐私陷阱会影响基于更强大模型的方法,这些模型表征人类心理的个人方面。Big-2 的有效性评估分为两部分:对特定预测任务的定量评估和基于对 Big-2 特征可以从在线行为中得出的不同方式的分析的定性评估,提出了指导此类工作的通用模板。定量结果表明,我们的轻量级模型在预测任务中可以达到或超过 Big-5 的性能,而定性结果表明基于对基本在线用户行为的观察来实现该模型是可行的。我们的主要结果是一个通用模型,可用于以合乎道德的方式表征在线平台用户的个性特征。我们提出的模型提供了一个有价值的工具,可以对在线个性进行有效且可解释的分析,避免收集不必要的用户数据,这可能会导致违反道德的可能性。

更新日期:2021-07-29
down
wechat
bug