A Domain Knowledge Based Method on Automatic Generation of Information Requirement Against Big Data

Increasing number of enterprises are mining Big Data for decision support, but they found Big Data not quite as usable as their own well organized data-bases. Useful information is hidden within Big Data, and could only be found by search, according to keywords, queries or other kinds of requirement express patterns. It is found that during decision making process, user information requirement highly depends on the decision task, that on processing tasks of a same type, requirements have similar content and scope. Based on this principle, a method is proposed to automatically generate information requirement against Big Data, by making use of a new kind of domain knowledge-the latent mapping relations among decision task types and information requirement models. It is proved by tests that, the method can improve efficiency of decision support information collection from Big Data.