Derive: finding semantic concepts with property-values from natural language text
John Cuzzola, Ebrahim Bagheri · Computer Science and Software Engineering · 2014
There is a vast quantity of data embedded in natural language text in the form of objects and property-values. Correctly identifying these objects and matching them to their appropriate property would have immediate applications for areas of information retrieval and semantic reasoning. We present Derive, a framework that can locate and match properties to objects. Initial experimentation have revealed favorable results but more work is necessary particularly with property-value discovery.