基于灰关联分析的V-MDAV算法研究

距离度量会影响微聚集算法的聚类效果，为了提高算法的隐私保护能力，采用灰关联分析中的均衡接近度替代V-MDAV算法中的欧氏距离度量记录间的距离，提出基于灰关联分析的V-MDAV算法，即V-GRAV（variable-size grey relation to average vector）算法。由于均衡接近度既包含灰关联度对整体接近性的测度，又具有均衡度对序列均衡性测度的特点，克服了欧氏距离受局部奇异值影响较大的问题，所以V-GRAV算法在保证信息损失与V-MDAV相近的同时，较大程度地降低隐私泄露风险，实验证明了算法的有效性。

Research on V-MDAV algorithm based on grey relational analysis

Distance measure can affect the clustering effect of micro aggregation algorithms, in order to improve the privacy preserving ability of the algorithm, the Euclidean distance in the V-MDAV algorithm were replaced by the balanced adjacent degree in grey relational analysis method to measure the distance between records. This paper proposed the V-MDAV algorithm based on grey correlation, called V-GRAV algorithm. The balanced adjacent degree included the characteristic of the measure of grey relational degree to the whole approximation and balanced degree to the sequence balanced degree, which could eliminate the point correlation tendency. It overcame the problem that the Euclidean distance was greatly influenced by the local singular value. Therefore, the V-GRAV algorithm can reduce the privacy disclosure risk while ensuring that the information loss is similar to V-MDAV algorithm. This experiments demonstrate that the algorithm is effective.