Muf: Tool For Knowledge Extraction And Knowledge Base Building

Muf is a tool for knowledge base (KB) building by extracting knowledge from texts. It is intended for cases when all documents have to be processed manually in order to ensure correctness of KB. Muf provides visual tools and some degree of mechanization to facilitate manual knowledge extraction and KB building. We also believe that manual processing of documents implies small number of documents and it further implies that KB use case is well defined. Well defined use case allows us to decide which knowledge is worth of extraction and which not. Reducing amount of extracted knowledge also leads to less complex structure of KB. This all makes extraction and KB building tasks even easier, KB is easier to understand and deploy. If some parts of KB happen to be incorrect, Muf is able to trace the corresponding knowledge down to the text and allow user to fix it. The work was done on Czech drug labels, but we believe that Muf can be used also for different languages as well as different kinds of documents.