具保護隱私之高效率資料比對協定

Abstract

資料比對可以判斷那一對資料，記錄著在真實世界中的同一個物體，它是在資料探勘領域中，一個重要並且基本的資料整合運作。近幾年由於對網路安全的意識抬頭，保護隱私已經引起廣大的注意。因此具保護隱私之資料比對協定的目的，就是去判定在兩個獨立資料庫中的共同資料紀錄，又同時避免洩露任何私密的資料，雖然保護隱私與資訊分享是兩個彼此衝突的概念。在本論文中，一組經改進用來設定資料轉換空間的參考集合被提出來，並由實驗顯示當為了保護資料私密而將明文資料值轉換到一度量空間時，它是一個更有效率且更精確的方法。

Record matching or data linkage identifies all pairs of matched records that refer to the same entity in the real world; it is an important issue and a basic operation of information integration in the data mining field. In recent years, preservation of privacy has gained a lot of attention because of an increasing awareness of the importance of security. Therefore, the aim of the privacy preserving record matching protocol is to recognize the common records shared between two autonomous data sources and keep privacy leakage low without revealing any private data at the same time, although the main idea of privacy-preserving and information sharing conflicts in nature. In this thesis, one modified reference set to set the embedding space is proposed and demonstrated by experiment that it is more efficient and precise when transforming plain record values into a metric space in order to keep confidential.