Induced Consent Analysis of Privacy Policy Based on Grounded Theory and Machine Learning
Analyzing privacy policies from the user's perspective to understand the tendency for induced consent is beneficial in helping users identify unfair terms and providing regulatory authorities with guidance to standardize app privacy policies.This study uses grounded theory to examine the tendency of induced consent in privacy policies from the user's perspective and develops a coding system for such tendencies.After manually annotating the corpus,we trained a K-BERT model using semi-supervised learning to achieve the automated identification of statements with a tendency to induce consent within privacy policies.Moreo-ver,further network analysis and sequence pattern mining were conducted to explore the characteristics and underlying patterns of user consent induction in privacy policies.Empirical analysis reveals that user opportu-nity costs,privacy management costs,and fuzzy concepts are central to the network of inducing dimensions.Fuzzy concepts and responsibility-shifting statements play a crucial role in the patterned inductive writing of privacy policies,usually appearing densely following other unfair statements.Furthermore,the study identi-fies significant differences in the features of induced consent between the children's domain and other do-mains.Some common features exist among privacy policies across specific domains,potentially linked to similarities in service delivery and business logic.
万方数据
维普
