Variables Extraction in Natural (English) Language Through Possessive Relationships

抄録

The already highlighted importance of the ‘flow’ of data in the Market of Data brings needs of development of ways to better ex- plore the utilization of data. Aware of the existence of rich knowledge stored and shared in text format, this paper aims to propose a form of representation of variable names that can be identified in natural language written knowledge. With the use of possessive relationships between words in Noun Phrases, we supported the representation of variable name relating a variable to a thing or event. A simple experiment was performed to demonstrate the efficacy of the proposed representation supported by Data Jacket Store, where we can find well-form variable names under the name of Variable Labels.