中文摘要：

解决大背景问题是当前的重要课题。给出了对象亏值、属性亏值、投影粒及共轭对的形式化表示方法、相关定理和证明。通过对象亏值和属性亏值获得原子共轭对，由交运算得到全部共轭对，提出了针对大背景概念格获得全部共轭对的一种新方法。该方法实现了按用户要求对概念格知识进行聚类化简，具备“无苛刻条件，化简幅度较大，运算过程简单，具有信息恢复能力，有明显对应关系”的优点。将其应用于中医的药材和处方，为中医专家提供决策支持。为共轭对在数据挖掘的深入应用奠定了基础。

英文摘要：

Solving the large context problem is currently an important subject. Object waned values, attribute waned values, projective particles and conjugate pairs are defined and the relevant theorems and proofs are given. The atomic conjugate pairs by object waned values and attribute waned values are gotten. All the conjugate pairs through intersection operation are obtained, and a new method of obtaining all the conjugate pairs for concept lattice from the large context is presented. The approach reali zes clustering reduction for concept lattice knowledge meeting user requirement and has the advantages of no harsh terms, with a wide reduction range, simple computing process, information recovery capacity, and distinct correspondence. It is applied to Chi nese medicinal materials and prescriptions, decision support is proyided for Chinese medical specialists, and the foundation is established for the in-depth applied research on data mining by conjugate pairs.