Quantification of perceived location privacy, and its relationship to privacy behaviour

隐私对于信息和通信技术来说是一个越来越重要的话题。从用户的在线活动和移动终端使用中收集大量数据，并将其聚合以形成可能影响用户隐私的信息。虽然存在控制对这些信息的访问的机制，但普通用户几乎不可能理解和控制可以从收集的数据中提取的信息。位置信息在这种情况下是特别关注的，因为从看似匿名的位置数据中可以推断出大量的信息。用户在使用基于位置的系统时的隐私偏好通常可以在二分的水平上进行测量：用户在特定情况下共享他们的位置，或者他们不共享。然而，用户的实际隐私偏好可能比这更细粒度。该研究项目的目标是（1）理解和量化这些偏好，（2）理解它们与公开共享行为的关系，以及（3）通过正式建模使这些知识可访问。第一个目标是通过量化值分配的方法来评估细粒度的隐私偏好。第二个目标是通过实证研究如何将这种量化的位置隐私评估反映在实际的隐私行为，考虑到用户的知识和意识的位置共享。第三个目标是通过调整现有的行为模式，使移动的环境中与隐私有关的行为正规化。

Privacy is an increasingly important topic for information and communication technologies. Massive amounts of data are collected from the users' online activities and mobile device usage, and aggregated to form information which can affect a user's privacy. Although mechanisms exist to control access to such information, it is mostly impossible for a common user to understand and control the information which can be extracted from collected data. Location information is of particular concern in this context, as from seemingly anonymous location data vast amounts of information can be inferred.Privacy preferences that users have when using location-based systems can typically be measured on a dichotomous level: either the users share their location in a certain situation or they do not. However, the users' actual privacy preferences are likely more fine-grained than this. It is the goal of this research project to (1) understand and quantify such preferences, (2) to understand how they relate to overt sharing behavior, and (3) to make this knowledge accessible through formal modelling. The first goal is reached by assessing fine-grained privacy preferences by a method of quantified value assignment. The second goal is reached by empirically studying how this quantified location privacy valuation is reflected in actual privacy behavior, taking the users' knowledge and awareness of location sharing into account. The third goal is reached by adapting existing behavioural models to formalize privacy-related behaviour in a mobile context.