| Summary: | 碩士 === 國立屏東科技大學 === 資訊管理系所 === 101 === K-anonymity is a method for data protection privacy, which lets all the data have at least k-1 item which is similar to the original data. This method can prevent the data from being identified by changing certain aspect of original data. Recently, among use of k-anonymity data clustering, data preprocessing before clustering, cluster formation method, selecting the start point of cluster and processing residual value of data, usually affect effectiveness of data mining after anonymizing. In view of this, this study proposes to k-anonymity based on glowworm evolution-based clustering algorithm(K-anonymity of Glowworm Swarm Optimization, abridge KGSO). At first, we collect good data from the dataset as a cluster starting point through the concept of glowworm swarm optimization. And we couple with the greedy k-member algorithm to establish a good way of successive clustering to distinguish between successive clustering and residual value processing stage, which can be further divided into the original k-member and revised k-member. These are two different ways of selecting the starting point to establish the cluster in the remaining value of the processing stage. The experimental result reveals that the KGSO-RK is better than the k-member on information loss, and improves the k-member quality of information in the anonymous effectiveness by reducing outlier effects, and thus to increase the effectiveness of post-mining.
|
|---|
