Chinese medical formulas are the main tool for clinical treatment. Chinese medical formulae have effectively improved prevention and treatment by Chinese medicines combination. However, as the basic substances，Chinese medicines （CM）are too complicated to make quality control. Also, the mechanism of TCM is still unknown. Many previous studies mainly focused on the experimental investigations，and the results are mostly effected by the experimental condition，technology and the research direction，it is hard to find the inner connections of formulae and Chinese medicines. Under the circumstances, information technology may give a great help for further research in TCM prescription compatibility by using the network analysis. Objective: Description prescription compatibility with visual network, and to analyze the relationship between the indexes of group’s network and property-flavor-meridian tropism (xing-wei-guijing), which were the three characters of TCM. The research aimed to find out the compatibility of TCM and also provide a reference of chinese medicine theory. Methods: First, colleting 4,827 prescriptions and 1570 herbs’ details from "TCM prescriptions database" and "TCM database" on the web of Shanghai Guidance of Science and Technology ".Second, deleted 153 prescriptions which compositions were lost. The ‘ Chinese medicines had been integrated according to "Chinese Materia Medica", "Chinese Dictionary alias (Amendment)".This research had included 4,674 prescriptions and 893 herbs’ details. Finally, the compatibility network of TCM prescriptions has been built up. Medicines were selected as a junction point for the network and their relationships were chosen on the sides. By using network analysis, it found out that the relationships between group’s network indexes and 澳門大學碩士學位論文 IV property-flavor-merdian tropism of TCM. Moreover, the internal relations would be verified by ANOVA and κ test. Result: According to the structure and properties of the network, it found that there are four major groups by using the network analysis, which were Community 1, Community 2, Community 3, Community 4, and the number of groups were respectively 247, 200, 311, 131. The internal of groups are closely related, higher frequency compatibility and the external of groups are converse. CM in community1 were mainly cold property (65.35 %), sweet(35.42 %)and the lung meridian (26.12 %), CM in Community2is mainly warm property (61.90 %), bitter flavor (35.75 %)and the liver meridian (30.93%), CM in Community3 is mainly warm property (70.60%), bitter flavor (34.18%)and the spleen meridian(19.40%),Community4 is mainly included cold property (64.79 %), bitter (36.84%)and the liver meridian(23.43 %).We can find that there were significant differences (P

中藥方劑是中醫臨床治療的主要工具，方劑通過中藥配伍提高了防病治病的 療效。中藥作為治療的物質基礎，其成分複雜，質量難以被控制，藥效機理尚不 十分明確。由於多數研究以實驗性探究為主，所得的結果受到實驗技術和條件的 牽制，加上各自研究角度的不同，使得研究結果較難呈現出系統性的聯繫。在此 背景下，借助資訊科技的力量，運用網絡分析對中藥方劑配伍進行更全面的研究 具有重要的意義。 目的：用可視化網絡描述中藥配伍，並分析相關網絡指標與中藥各藥性指標 間的關係，為中藥方劑配伍的解析提供參考。 方法：首先，在上海研發公共服務平臺之“中醫方劑數據庫”和“中藥數據庫” 裡，提取與審定方相關的 4827 條方劑和 1570 味中藥。刪除 153 條方劑組成缺失 的數據，再根據《中華本草》、《中藥別名大辭典(修訂本)》對所提取的中藥名 稱進行整合，最終收錄了 4674 條方劑和 893 味中藥。然後，以中藥為結點、中 藥配伍的關係為邊，構建中藥方劑配伍網絡。運用網絡分析尋找群體網絡指標與 性味歸經等中藥指標之間的關係，並用方差分析和卡方檢驗等統計學方法進一步 驗證。 結果：根據網絡的結構與屬性，通過網絡分析發現該網絡主要有四大群體， 依次為群體 1、群體 2、群體 3、群體 4，四大群體所包含的中藥數目分別為 247、 200、311、131，群體內部各中藥彼此間關係密切、配伍頻次較多，而群體之間 的中藥則聯繫較為稀疏、配伍頻次較少。群體 1 裡中藥以寒性(65.35 %)、甘味 (35.42 %)為主，多歸肺經(26.12 %)；群體 2 裡中藥以溫性(61.90 %)、苦味(35.75 %) 為主，多歸肝經(30.93 %)；群體 3 裡中藥以溫性(70.66 %)、甘味(34.18 %)為主， 主歸脾經(19.40 %)；群體 4 裡中藥以寒性(64.79 %)、苦味(36.84 %)為主，多歸肝 經(23.43 %)。通過統計學分析發現：①不同網絡群體的中藥在性、味、歸經方面 澳門大學碩士學位論文 VII 均存在差異顯著性(P<0.05)；③中藥的歸經與間隔中心度的差異存在顯著性(P