Abstract
Finding technologies in the text of patents or other documents such as medical articles is a subtask of building a technology ontology. Building such a technology ontology was proposed by Brandeis scholars as part of the project aimed at patent classification based on a certain notion of availability of technologies relevant to the patent. Technology ontology represents a database of technologies evaluated by their availability within certain time frame, that is their maturity. Technology terms identification in the text of documents is an initial step necessary for building an ontology. The terms found in the text of the patent will reflect the notion of a technology and constitute the basis for technology maturity identification.\r \r \r Here, we explore the efficiency of using natural language processing techniques to help identify technologies in patent text. We attempt at creating and using a matcher that uses lexical and syntactic features to look for technology terms. We address the problem of determining the concept of a technology which is important for the task and use an annotation for the evaluation of the matcher. Finally, we analyze the results and propose improvements to the system.