混合值差度量在MDS算法中的应用

Application of heterogeneous value difference metric on MDS algorithm

多维尺度分析(MDS)通常以欧氏空间中点的距离来度量对象间的差异性(相似性)。当对象有像性别、颜色等名义属性时,通常的做法是将它们数量化,然后再对其运用欧氏距离,显然,这种处理方法存在不合理性。将一种混合值差度量(HVDM)引入含名义属性的对象间距离的计算,以改善名义属性下MDS的计算合理性。在UCI Abalone数据集上进行的实验,结果表明该方法比传统的数量化方法在重构能力、重构精确度方面都有更好的表现。

In general,Multidimensional Scaling（MDS） uses Euclidean distance to measure the dissimilarity（similarity） of objects.If objects have nominal attributes,such as sex or color,common practice is digitizing first and then applying Euclidean distance.Obviously,this approach is not reasonable to some extents.The Heterogeneous Value Difference Metric（HVDM）,a distance metric computing distance for nominal attributes differently than Euclidean distance,is applied to MDS to improve its reasonableness on nominal attributes.Experimental results on UCI Abalone dataset show that the proposed method gives promising results on both reconstruction ability and accuracy.