Creating a patient data base using the problem-directed semantic net
The Problem-Directed Semantic Net as proposed by Jordan is utilized in the computer inplementation of a medical records system. Natural language input is accepted by the system and subjected to a partial syntactic parse. Medically based case analysis routines then build case structures representing the meaning of the input string. These structures are stored in semantic networks. The resulting data base can be subsequently used for many different types of retrieval, such as a patient history, one particular visit of a patient, or a group of patients with the same problem.
The initial implementation of this system accepts only simple sentences and is limited to the area of the physical examination. The modular programming approach used in this system would facilitate its extention to complex sentences and to other areas of medical discourse.
Thesis78E356.pdf
3.32 MB
Adobe PDF
1c29c613923882bd2ca9bd531f439ecd