A computational approach for processing locative and temporal information in clinical medical records /
| Main Author: | |
|---|---|
| Other Authors: | , , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1989.
|
| Subjects: | |
| Online Access: | Link to OAKTrust copy |
| Abstract: | This research designed and developed a methodology for processing unrestricted natural language to obtain a specific type of information. The motivation for this project came from medical research that relies on epidemiological studies conducted on a large set of clinical records collected from physicians recording individual patient observations. Among other information, these records contain data about the geographical (locative) concentration of diseases. Processing of clinical records must deal with temporal, locative, and conceptual issues. This makes text understanding and data extraction of clinical records an excellent area for applied research. While much has been done in making temporal or conceptual inferences in medical texts, parallel work in locative inferences has not been done. This dissertation examines the locative inferences that are part of both the larger problem of the integration of temporal, locative, and conceptual issues in the clinical record understanding domain as well as the development of a methodology for use across other domains. In this dissertation, a prototype demonstrating the concepts of this methodology was developed. This prototype, the Intelligent Data Collection System (IDCS), demonstrated (1) necessary preprocessing of unrestricted natural language, (2) parsing specifically looking for locative and temporal information, (3) coordinating locative and temporal information on a time-graph, and (4) making common sense inferences about the completeness of the data. |
|---|---|
| Item Description: | Typescript (photocopy). Vita. "Major subject: Computer Science." |
| Physical Description: | xi, 124 leaves : illustrations ; 29 cm |
| Bibliography: | Includes bibliographical references. |