In large data space, data process is simplified by using cosine similarity and some special vectors are got, these vectors are response to the changing of elements, by which the vectors are composed.

• 在对超大数据空间进行数据处理时,用余弦相似度聚类分析,得到一些反映向量对应元素之间变化的特征向量,用这些向量作为因变量空间代替原数据空间,再用偏最小二乘法对其进行关联,是一种快速有效的方法。