Study On Approximate Reasoning Methods Based On Formal Context

Concept lattice is an important model for data processing and analysis. Study on the re-lationship between objects and attributes of a concept lattice is a concrete reflection of humanmind reflecting the special relationship among concepts. The relationship between some objectsand attributes is fuzzy in nature, which can generate a lot of fuzzy concepts, fuzzy decisionsand fuzzy classifications. Fuzzy concept lattice has become an important topic of artificialintelligence area.There is a lot of uncertain phenomena in real life. In recent years, approximate reasoningbased on uncertain problems has been widely concerned and become a hot study field of artifi-cial intelligence. This dissertation introduces approximate reasoning methods to formal contextand establishes a relatively intuitive relationship model between condition attributes and de-cision attributes, reflecting the process of reasoning and realizing data mining and knowledgediscovery in decision formal context, thus guiding people’s behavior scientifically.The main work of this dissertation can be generalized as follows:1. This dissertation introduces subjective Bayesian probabilistic reasoning method to de-cision formal context and discusses the relationship between condition attributes and decisionattributes. Inclusion degree is used in the reasoning, thus finding a subjective Bayesian proba-bilistic reasoning method without a priori probability.2. Fuzzy concept lattice based on Goguen implication operator is constructed, then decision is proposed between condition attributes and decision attributes about this kind offuzzy concept lattice. Finally, a study on approximate reasoning is discussed about this conceptlattice.3. Fuzzy decision implication is introduced to fuzzy decision formal context, then we dis-cuss the semantic interpretation of fuzzy decision implication.