查詢結果分析
來源資料
相關文獻
- 以類分子神經系統對B型肝炎臨床資料作診斷預測、分析、及交叉驗証
- 慢性B型肝炎之治療
- Immunogenicity and Reactogenicity of Two Recombinant Hepatitis B Vaccines in Healthy Adolescents on Two-dose Schedule
- B型肝炎帶原的中醫療法
- 醫療尖銳物品扎傷追蹤調查
- Reactive Hemophagocytic Syndrome in Patient with Acute Hepatitis B: Report of a Case
- 免洗餐具的誕生:醫學知識在臺灣的社會性格分析
- Control of Hepatitis B Virus Infection by Vaccination: The Taiwan Experience
- Immunopathogenesis of Viral Hepatitis B and C
- 慢性B型肝炎病毒感染的自發性發作與緩解的病因及臨床意義
頁籤選單縮合
題 名 | 以類分子神經系統對B型肝炎臨床資料作診斷預測、分析、及交叉驗証=Diagnosis, Analysis and Cross Verification of Clinical Hepatitis-B Data with an Artificial Neuromolecular System |
---|---|
作 者 | 陳重臣; 方國定; | 書刊名 | 管理與系統 |
卷 期 | 6:4 1999.10[民88.10] |
頁 次 | 頁433-457 |
分類號 | 415.5332 |
關鍵詞 | B型肝炎; 資料區別; 自主性學習; 逐漸轉變; Hepatitis-B; Data differentiation; Self-organizing learning; Gradual transformability; |
語 文 | 中文(Chinese) |
中文摘要 | 本文主要是利用一個動機起源於已知大腦神經生理學方面的研究,加上 一些假設,而建立的類分子神經系統,應用在B型慢性肝炎的診斷分析上。它與 一般傳統類神經網路最大不同的地方,是整合神經元內部及神經元間不同方式的 資訊處理,以取得類生物體結構/功能逐漸轉變的特性,即當系統結構產生些微 改變時,它的功能也跟著產生些微的改變。系統對676筆B型肝炎檢查資料(1/2是 B型肝炎患者,1/2是健康者)作區別分析學習,每筆資料有十個檢查項目。經過 學習後,系統對資料的區別(認識)比率高達99.3,然後,以另一組375筆B型肝炎 檢查資料作測試,結果顯示,系統判斷受測者是否具有肝炎的正確率高達95.7。 系統並以自我調整學習的方式,從受測者的資料型態中,決定重要及不重要的B 型肝炎檢查項目。另一結果顯示,當Alb/Glo比值越小或GPT/GOT越大,受測者罹 患B型肝炎的可能性相對較高。 |
英文摘要 | This paper described the application of an artificial neuromolecular system, motivated from some physiological evidences along with some hypotheses, to the diagnosis of hepatitis-B clinical data. Unlike traditional neural networks, the system integrates different modes of intra- and inter-neuronal information processing to capture the biological feature of gradual transformability of structures/functions. That is, the fuctions of a system alter gradually as its structures change slightly. The system was tested with a database of 676 patients, of which one half are hepatitis-B positive and the other are negative, each consisting of 10 clinical test items. Experimental results showed that the system is able to correctly differentiate 99.3 of these 676 patients' clinical data. The system after substantial learning was then tested with another database of 375 patients, and showed that it is able to correctly differentiate 95.7 of the 375 patients' clinical data. From the structures of the clinical data, the system is able to determine significant and insignificnt test items through self-organzing learning. It also showed that the chance of being infected with hepatitis-B is high when a patient has either a low Alb/Glo rate or a high OPT/GOT rate. |
本系統中英文摘要資訊取自各篇刊載內容。