A fast batch algorithm for concept generating based on concept dictionary

The efficiency of concept lattice construction is a prevalent topic because it is necessary and important to construct concept lattice in many application fields concerning concept learning. This paper suggests a novel batch algorithm referring to the idea of isomorphic generating and closure operator. After context preprocessing in which the value and weight of objects are figured out, our algorithm could quickly generate all the possible concepts added by an object of context through using concept dictionary. Then we could obtain all the concepts after removing pseudo concepts from concept set. Our algorithm can reduce the search space when generating direct partial order relation easily between all concepts. Finally, our algorithm is completed as well as NextClosure algorithm using C# on windows platform, and comparison was made between them in the some data sets. Results showed that our algorithm performed better than NextClosure algorithm on temporal aspect.