Abstract: The theories of reduction of formal context in formal concept analysis are studied in this paper and the method of attribute reduction is applied to research of data cube. The base table can be reduced through removing unnecessary attributes and combining relative necessary attributes. Research shows that the constructive processing of data cube deduced from reductive table is easier than that from original table and data cube deduced from reductive table can obtain same aggregate concepts by storing process of reduction. After reduction of base table, the different roles of attributes played by different kinds of attributes in data cube are clear which can help us understand data characteristics deeply.

Key words: formal context, attribute reduction, data cube

摘要： 研究形式背景约简理论，将形式概念分析理论的属性约简方法应用到数据立方体的研究，通过对绝对不必要属性和相对必要属性约简实现基本表的简化。研究表明，由约简后的基本表进行的数据立方体的构建要比由原始基本表的构建过程简单，由于保存了约简过程，可得到与原基本表相同的聚集概念。通过对基本表进行约简，还明确了基本表中各个属性对数据立方体聚集单元的刻画作用，这对于理解数据的特性有重要的意义。

关键词: 形式背景, 属性约简, 数据立方体