A bi-directional fuzzy logic theory: the generalized Knuth's triadic logic for information retrieval

摘要

Standard logic only deals with true and false, i.e. {0, 1}. Zadeb discovered a type of logic that can contain values in 10, 1J which is called fuzzy logic today. On the other hand, the many-valued logic was first studied by Lukasiewicz. He created a so-called triadic logic by adding a 112 into the binary set {0, 1}. Knuth modified Lukasiewicz's triadic logic to {-1, 0, +1} rather than Lukasiewicz's {0, 1/2, 1}. This paper will combine Fuzzy logic with Knuth's {-1, 0, +1}, and focus on developing a bidirectional fuzzy logic theory applied to information retrieval with expert relevance feedback and document relevance computing.