| Summary: | 碩士 === 國立臺南大學 === 數位學習科技學系碩士班 === 96 === As the recent development of the computer science, the data quantity of enterprise database increases rapidly. To extract the usefulness information from huge databases, many efficient data mining technologies have been applied. In recent years, the data mining tools are more and more powerful, and the risk of privacy leak has become an urgent problem. Privacy preserving data mining is a relatively new research area in data mining and knowledge discovery. In a common situation, databases are distributed among several organizations who would like to cooperate mining to extract global knowledge, but each party needs prevent it’s privacy not directly sharing the data. Therefore, this study presents an algorithm for privacy preserving distributed decision tree based on C4.5. While this has been done for horizontally partitioned data, this study presents an algorithm for vertically partitioned attributes. Each site computes a portion of data, and then they exchange the result to each other. The goal of this paper is to obtain correct data mining results and preserve the privacy of each site.
|
|---|
