Integrating multi-source user data to enhance privacy in social interaction

Building future generation service-oriented information broker networks

“…The implementation of the di.me platform prototype technologically enables personal data usage in a controlled, trustworthy, and intelligent way [14]. It specifies a platform incorporating user-control deeply in design: a personal server (PS) that enables a di.me node in a decentralized network to connect to other users' PSs or external services, like various social networking platforms as mentioned above, and this by using distinct identities [15].…”

“…This node integrates all personal data in a personal information sphere, including user interests, contact information, files or resources, and social network services. Intelligent features and PETs further guide user interactions with the digital sphere, illustrated by context-aware access control [16], trust and privacy advice, or organizing their personal information sphere [14]. Besides integrating existing networks and services, the platform provides its own OSN functionalities, which are not available in known and popular OSN, in particular network anonymity [17][18] [19].…”

“…Its high-level PS internal architecture, shown in Fig. 4, is related to that of dynamic webapplications [14], and also favored by the separation of the addressed concerns inherent to the multilayersystems. Within this approach, the persistence layer isolation also benefits the decoupling from the underlying database technology and enables a multi-engine deployment.…”

“…on Twitter through di.me, di.me's semantic analysis of the message recognizes an activity (sitting), a location check-in (beach), and a person (Carol). Triggers set by the controller module in the datamining module are fired, as the combination of personal information relating to third parties (Carol's identity) and their common location (beach) is a potential privacy issue, and this causes di.me to present Alice with a warning message informing her about that risk, and asking if she is sure she wants to post [14]. Unlike popular applications where users are expected to have these digital literacy skills, di.me allows non-literate users to participate in online social networks while informing them of risks only during relevant situations and thus minimizing the likelihood that the warnings get ignored.…”

“…Trust metric [14][16][28] [29]: An adaptive user trust index allows warning messages to be displayed, preventing a user from unknowingly sending a confidential file to the wrong audience. This concept is also applicablebeyond the di.me prototype no interactions in and between social circles, such as friends or business contacts.…”