Evaluación de prácticas de privacidad en aplicaciones móviles: desarrollo de un módulo de etiquetado de prácticas de transferencia de datos personales en políticas de privacidad en español usando técnicas PLN y aprendizaje automático.

Abstract

Failing to comply with requirements set out in the Ley Orgánica de Protección de Datos Personales (LOPDP) may lead penalties to the involved organizations. There exist transparency-related requirements, which seek to inform users on personal data processing. Privacy policies are the main mechanism to provide such a transparency. To assess compliance with statements laid down in privacy policies, privacy practices firstly need to be extracted. Yet, such extraction cannot be manually conducted, since it would not scale in the current digital ecosystem, which have a vast number of devices (and policies). To continue with the efforts to contribute to the provision of techniques and tools for compliance assessment with privacy and data protection requirements, this work presents a module for tagging data transfer practices in Spanish privacy policies, using machine learning and natural language processing algorithms. Chapter 1 presents the theoretical background related to privacy and data protection, and to machine learning and natural language processing techniques, which are necessary to understand this document. Furthermore, it details the tools and programming languages used in this work. Chapter 2 presents the methodology used to develop the module for tagging data transfer practices in Spanish privacy policies. In addition, the codes used to build the machine learningbased classifiers, to perform tests and to determine the models with the best performance are presented. An ensemble is built with the selected models and standard metrics are calculated to evaluate the performance of the elaborated models. Finally, Chapter 3 analyses the results obtained during the process and presents the conclusions and recommendations once the project has been completed.