以知識本體技術塑模會計知識之研究

Using Ontological Technologies to Model Accounting Knowledge

本研究利用本體（Ontology）技術發展一個知識庫系統，用於解決會計領域的專業問題，系統開發參考TOVE程序，將會計知識建置為階層架構及正規語意描述的領域本體，實務問題則以任務本體表達。本研究以IFRS和ROCGAAP對長期工程合約之規範為例，探討會計處理、交易認列判斷條件上之異同，並以測試交易案例來驗證知識推論。主要設計包括：(1)釐清議題的知識內涵；(2)將通用知識建置為領域本體；(3)將會計實務發展為任務本體；(4)利用推論機制推導隱含性知識。簡言之，本研究以本體為方法論、OWL為知識表達語言、Protégé為本體開發的工具。由實驗結果顯示：本研究已初步實現會計知識轉移至資訊系統，達成知識再利用的目標，研究成果將可提供未來建置智慧化會計系統的參考。

This study utilizes ontological techniques to develop a knowledge model and a knowledge base system for the domain of accounting profession. It uses TOVE (TOronto virtual enterprise) to develop an accounting domain taxonomy and formal semantics for the expression of the content knowledge. The “long-term construction contract” defined in IFRS and ROC GAAP is a case example for demonstrating the consistencies and differences in accounting processing and transaction recognition. The research design includes: (1) clarifying the knowledge content; (2) building a domain ontology based on general-purpose knowledge; (3) creating a task ontology based on a goal-oriented and problem-solving knowledge; and (4) using inference mechanisms to obtain implicit knowledge. In short, this study uses ontology as the methodology, OWL as the language of knowledge representation, and Protégé as the editing tool for developing ontology. The experiment results show that the knowledge classification and integration developed in this study have preliminarily achieved the objectives of human-machine knowledge transfer, domain knowledge sharing and reuse. The results of this study also can provide reference for building next generation accounting systems.