以本體論為基礎的情境感知隱私保護架構

Abstract

層出不窮的資料外洩新聞，打擊了民眾的使用意願。研究指出，成功的電子商務有賴於妥善保護個人隱私。W3C制定的P3P及APPEL幫助網站公布隱私政策，讓使用者自訂隱私偏好。也有學者提出不同的平台來保護私密資訊。然而，現有的體系皆未考量使用者在不同情境下會有不同的隱私偏好，以及隱私偏好設定之繁複；演算架構亦不適用於未來無所不在的運算環境。本研究提出以本體論為基礎的平台，希望賦予使用者更精確的隱私控制能力，並減輕設定上的困擾，以適用於快速改變的、充滿大量情境資訊的環境。 本研究運用W3C的語意網標準語言OWL，讓使用者依情境定義不同精確程度的隱私偏好。根據使用者設定的隱私偏好以及不同的情境資訊，本研究導入描述性邏輯的演算架構，用以自動決定資訊是否可以公開，或是否要接受提供者傳遞的資訊。本研究並提出混合式調適演算法，針對無所不在的運算環境中，情境資訊快速變動的情況，改善現有研究在此環境中的效能缺失。 透過知識本體驅動平台，使用者被賦予多層次且豐富的隱私偏好表達能力，並降低使用者被無所不在的服務要求淹沒的可能性。此技術使系統可以根據使用者的真實偏好，做出正確決定，減輕使用者對於隱私侵害的顧慮，增進服務被接受的可能性。

The incessant news of data security breach has reduced the willingness of public to use the Internet network. Research has shown that successful e-Commerce relies heavily on effective protection of privacy. The launch of P3P and APPEL by W3C has assisted websites in announcing their privacy policies and allowed users to customize their privacy preference. Some scholars, on the other hand, have proposed various platforms to protect private information. However, the current systems takes into account neither users’ privacy preference under different contexts nor the complexity of privacy preference customization; and the algorithm framework may not fit the ubiquitous computing environment in the future. This study proposes an ontological platform that empowers users with higher precision control over their personal privacy customization and alleviates the difficulties faced in setting, so that it is suitable for a fast changing environment filled with vast quantity of contextual information. This study adopts the semantic web standard language OWL, provided by W3C, to let users define privacy preference at varied accuracy levels under different contexts. Based on user-customized privacy preference and diverse contextual information, this study introduces a descriptive logic algorithm framework to automatically determine whether to disclose personal information or accept information transmitted by providers. In addition, this study utilizes Hybrid Adaptive Reasoning Algorithm to improve research efficiency on Ubiquitous Computing Environment where contextual information changes rapidly. The Ontology Driven Platform offers users multi-dimensional expression capability with privacy preference and reduces the possibility of being overwhelmed by ubiquitous service requests. The technology enables the system to make correct decisions in accordance with users’ genuine preferences, alleviate concerns about privacy invasion, and thus increase service acceptance.