| Summary: | 博士 === 國立臺灣大學 === 資訊管理研究所 === 91 === As the advances of information technologies, more and more
companies are willing to provide personalized services in order to enhance their customers'' satisfaction and loyalty. Consequently, people may use several personalized services provided by different companies, which may have their own database about personal data and collect their customers'' data independently. Under these circumstances, the concept of a universal profile system (UPS) has emerged. A universal profile system intends to provide a shared and integrated source (or storage) of personal data among a large number of service providers to bring the following benefits: (1) Services can project a person from a broader aspect, and so better quality of service can be provided. (2) A person''s switching cost among different services can be considerably reduced because different services can share the same copy of personal data. (3) A person only need to maintain one copy of personal data --- his universal profile, rather than inputting these data over and over again at different places and maintaining consistency among them. However, the benefits of a universal profile system cannot be realized. It is because a company may not be reluctant to share its customers'' data with others. Moreover, a
person may not wish to reveal details of his personal data directly.
In light of the obstacles, we design and implement a prototype of a universal profile system, called Personal Data Backbone (PDB) to provide foundation of a realizable universal profile
system in the following directions: (1) To ensure that the use of a person''s data is strictly under the person''s consent with a
personal data protection framework, called Online Personal Data Licensing (OPDL) (2) To let a user enjoy the benefits of personalization without revealing the details of personal data
with a mechanism of Derived Data Service (DDS). DDS can be viewed as an abstraction mechanism that provides Web-based
services with high level descriptions of a user''s characteristics, but hides the detailed of the user''s behavior. (3) To encourage a service provider to join PDB, which is provided by individuals themselves over peer-to-peer network.
|
|---|
