擴充性獸醫線上診斷專家系統之開發(以犬眼疾病診斷為例)

Abstract

目前台灣將專家系統運用在醫療診斷上的研究，大多以醫學為主，在獸醫學方面較少。有鑑於此，本研究希望開發一套可擴充性的獸醫線上診斷專家系統，提供獸醫師在臨床診斷方面準確且有效的協助。在程式設計方面，採用角色存取控管，將使用者分為系統管理員、獸醫師、一般使用者等三種角色，根據不同角色之功能需求給予不同的操作權限，以達到系統控管之目的。系統操作步驟如下，使用者首先要選擇動物的品種，接著選取出現的症狀，系統便會依據被選取的症狀，先經過法則式知識庫篩選，再透過加權計算後產生結果，告知使用者得到哪些疾病的可能性較大。本研究與獸醫系合作，由獸醫專家協助設定權重及資料庫的建立，首先針對犬的眼科疾病建立線上診斷專家系統，最終目的是希望藉由此系統所建立之診斷專家系統框架，透過獸醫師的配合，逐漸擴充資料庫，將系統的應用範圍延伸到其他動物疾病上，及經由持續更新權重資料，提高系統診斷的正確性，進而提供獸醫師在臨床診斷上準確且有效的協助。對一般使用者而言，也能透過這套系統，學習動物疾病的相關知識。隨著寬頻網路的普及化，線上診斷專家系統可以讓身處世界各地的眾多使用者，同時使用這套系統，不會受到時間跟空間的限制，配合友善的使用者介面，相信對這套系統的延伸性跟推廣更有助益。

The studies in the expert system of medical diagnoses in Taiwan are mostly on human diseases, very few are on animal diseases. In view of this, it is desired to develop an expandable web-based veterinary diagnosis expert system which will provide accurate and reliable assistance in clinical diagnosis for veterinarians. This system is designed in Role-Based Access Control (RBAC). There are three types of roles including administrator, veterinarian, and common user. Based on the needs of the role, each role has different authority operating this system. The main operation workflow is as fallow: First, user has to choose the animal’s category, than choose the symptoms appeared on the animal. According to the symptoms input by user, the system will search from the rule-based database and diagnose diseases by using weighting method. This expert system will notify user the possibilities of the diseases the animal might be infected with and refer to the recommended clinical diagnosis. This study is cooperated with the Department of Veterinary Medicine. The veterinary experts helped us define the weighting factor and created the animal disease database in this study. It is initiated from canine ophthalmology online diagnostic expert system then can be expanded to other animal diseases by veterinarians through the expert system frame developed. Via continuous updates of weighting factor to increase the accuracy of diagnosis will provide accurate and reliable assistance in clinical diagnosis for veterinarians. Additionally, common users can learn the knowledge of animal disease through this system. Worldwide users can operate this system at the same time, with no time frame limitation by utilizing the universality of broadband network.