CAJ | 학술논문

实例是本体的重要组成部分，它在很大程度上决定了本体的可用性。而目前本体实例构建的难度甚至超过了本体构建本身，大多实例的获取、更新和扩充依靠人工完成，既花费大量时间，又难以保证质量。文章在已完成的水稻本体概念框架基础上，利用神经网络方法进行半自动水稻实例抽取，提出水稻本体实例构建框架。统计数据表明，该方法能够有效地提高本体实例构建效率，大幅度降低手工劳动水平，提高本体实例质量，为本体实例构建和本体走向实际应用提供了思路和方法。
실례시본체적중요조성부분，타재흔대정도상결정료본체적가용성。이목전본체실례구건적난도심지초과료본체구건본신，대다실례적획취、경신화확충의고인공완성，기화비대량시간，우난이보증질량。문장재이완성적수도본체개념광가기출상，이용신경망락방법진행반자동수도실례추취，제출수도본체실례구건광가。통계수거표명，해방법능구유효지제고본체실례구건효솔，대폭도강저수공노동수평，제고본체실례질량，위본체실례구건화본체주향실제응용제공료사로화방법。
Instance is an important component of Ontology, which largely determines the availability of Ontology. Constructing an Ontology Instance is more difficult than Ontology construction itself in many cases. Instance acquiring, updating and expanding rely on manual operation completely. The work needs a lot of time and is hard to ensure quality. Based on the conceptual framework of Ontology, a semi-automatic extraction approach was applied by using neural network (NN) method in rice instances. The construction framework of rice ontology instance was proposed. The result shows that this approach effectively improved the efficiency of Ontology instance construction, significantly reduced the manual labor, and improved the quality of Ontology Instances. This approach provides a new idea and method for Ontology applications from Ontology Instance construction and Ontology.