|
Vol: 58(72) No: 1 / March 2013 Code Automation and Data Virtualization Based on a Natural Language Interpreter Used for Medical Software Development Florin Cumpănau Babes Bolyai University, Faculty of Mathematics and Computer Sciences, Str. M Kogalnicieanu Nr. 1, 400084 Cluj Napoca, Romania, phone: (+40) 264-405-327, e-mail: office@profivet.ro Cosmin Mureşan Emergency hospital, University of Agricultural Sciences and Veterinary Medicine, Faculty of Veterinary Medicine, Calea Manastur nr. 3-5, 400372 Cluj Napoca, Romania, e-mail: office@profivet.ro Nicuşor Lare Babes Bolyai University, Faculty of Mathematics and Computer Sciences, Str. M Kogalnicieanu Nr. 1, 400084 Cluj Napoca, Romania Bogdan Ionuţ Şanta Babes Bolyai University, Faculty of Mathematics and Computer Sciences, Str. M Kogalnicieanu Nr. 1, 400084 Cluj Napoca, Romania Keywords: natural language, medical application, language conversion, software automation Abstract In order to create a software development environment shared by physicians and computer scientists we defined a restricted natural language and created a natural language interpreter and a code automation utility. The code automation is used to extend the application with new objects described by the user. One of our goal and the reason why we implemented the code automation is to develop a basic reverse engineering for medical texts. During the development we considered both computer scientists and physicians expectations. We are currently at the forth application that accept natural language input. Each application was created with different programming technologies and the natural language interpreter was redesigned to accept new features. Major changes were made regarding the internal representation of objects and relations in order to facilitate data persistency and natural to programming language conversion. Various model explorers have been created to debug natural language applications. The current state of development allows physicians to describe their working environment in natural language. As a result a basic application is generated containing the data model as it was described by the physician in natural language. The application provides a graphical model explorer with data persistency capabilities. Simulation, conditional control and context validators are intended to be added to the existent capabilities. Rebuilding the natural interpreter with various technologies and testing each implementation resulted in a better understanding of the role of physicians in the medical software development. Many expected capabilities are still missing from the actual medical applications created around the restricted natural language interpreter and less restriction are intended to be used in the next future. References [1] A. Rodney Brooks, “Intelligence without representation”, Artificial Intelligence Volume 47, Issues 1-3, D. Reidel Publishing Company, January 1991, pp. 139-159 [2] R. L. Warda and W.V Huang, “Simulation with Object Oriented Programming”, Computers Industrial Engineering Volume 23, Issues 1-4, Elsevier Ltd. Publishing Company, November 1992, pp. 219-222 [3] V. Gustina and J. Viranta, “Pattern recognition with fuzzy neural network”, Microprocessing and Microprogramming, Volume 40, Issues 10-12 Elsevier B.V. Publishing Company, December 1994, pp. 935-938 [4] S.W. Lokea and A. Davison, “Pattern recognition with fuzzy neural network”, The Journal of Logic Programming Volume 36, Issue 3, , Elsevier Science Inc. Publishing Company, September 1998, pp. 195-240 [5] I. Popescu, “A relational model for knowledge representation in expert systems”, Journal of Systems and Software, Volume 18, Issue 2, Elsevier Inc. Publishing Company, May 1992, pp. 147-155 [6] F. Cumpănaşu, C. Mureşan, N. Lare and B. I. Şanta, “Applications of a restricted natural language for medical software development without programming knowledge”, Proceedings of 32nd National Conference on Medical Informatics (RO-MEDINF 2012), Timisoara, Romania, 2012. |