中国科学院软件研究所机构知识库
Advanced  
ISCAS OpenIR  > 中科院软件所  > 中科院软件所
题名:
基于本体的军事知识联通
作者: 雷玉霞
答辩日期: 2002
专业: 基础数学
授予单位: 中国科学院软件研究所
授予地点: 中国科学院软件研究所
学位: 博士
关键词: 军事知识表示 ; 军事知识获取 ; 军事本体 ; 军事知识分析 ; 军事知识联通
其他题名: Ontology-Based Interconnection of Military Knowledge
摘要: 军事专业知识在军事决策系统、军事智能教学系统、军事知识查询系统等高技术产品中具有重要的作用.知识库是机器智能的重要来源,所以有必要建立一个完善、协调、具有良好联通性的大型知识库,而大量实践证明计算机真正具有智能不仅需要大型知识库,而且更需要知识间广泛的联通.所谓知识联通就是通过某种途径将两条知识联系起来的过程.我们把知识联通分为三类:概念联通,陈述联通和本体联通.该文主要讨论概念联通.作为在知识服务方面的应用,在第5节给出了两个联通算法.第1个联通算法的设计目的是:输入两个概念C<,1>和C<,2>,尽可能多地输出有关C<,1>和C<,2>的有意义的知识.设计思路为:首先考虑C<,1>和C<,2>之间的关系联通,然后考虑C<,1>和C<,2>之间的属性联通,最后考虑C<,1>和C<,2>之间的属性值联通.第2个算法的设计目的是:输出两个概念C<,1>和C<,2>的属性值的定量比较.其设计思路是:将服从特殊概率分布的属性看成随机变量,然后求出该随机变量的期望,最后比较两个随机变量的期望.军事概念联通的意义主要体现在知识获取阶段和军事知识服务系统(例如自然语言查询系统)中.
英文摘要: Many applications need sufficient background knowledge of military domain. Typical applications include military decision-making systems, intelligent tutoring systems, and military knowledge query systems. It is necessary for us to build a large knowledge base as the back ground knowledge, which is as correct and robust as possible. Moreover, facts have proved that it is necessary for computer to have intelligence, i.e., have enough knowledge and interconnections among knowledge. Knowledge interconnection is to build certain connections between two pieces of knowledge in terms of certain methods. The knowledge interconnection is decomposed into three classes: conceptual, statemental, and ontological. In this paper, we mainly consider conceptual, which focus on some connections between two concepts via the relation-links or attributes or values of attributes. A relation-link is a sequence of the form "C_1,R_1,C_2,R_2…R_(n-1), C_n", where Q is a concept, R_j is a relation, l≤i≤n, l≤j≤n-l. Conceptual can help us to find inconsistent knowledge in a large knowledge base. The concept interconnection has three kinds of forms: interconnection in fixed classification H, attribute-based interconnection, and relation-link-based interconnection. The relation-link-based interconnection is similar to semantic connection in MindNet. We define the least upper-concept lub(C,C) of C and C as the least concept C" such that C,C'≤C" and C"≤D for any C,C'≤D. We define the great lower-concept glb(C,C) of C and C as the great concept such that C"≤C,C' and D≤C" for any D≤C,C. Our measure of interconnection-degree in H satisfies the following intuitive properties: The more two concepts have the least upper-concepts(the great lower-concepts), the bigger the interconnection-degree is. Interconnection-degree of two concepts having both the least-concepts and the great lower-concepts should be stronger than that of concepts having only one. If the paths of two pairs of concepts are symmetric in H, then they have the same interconnection-degree. The smaller the difference of levels in H is, the bigger the interconnection-degree of two concepts is. In order to discuss attribute-based interconnection, we differentiate structural-attributes, taxonomical attributes and unstructured attributes according to the domains of attributes. An attribute is structural if its domain H is structural, that is, there is a certain partial order over H. An attribute is taxonomical if its domain is taxonomical. An attribute is unstructured if its domain is unstructured. Our measure of the interconnection-degree of two concepts also satisfies the properties as follows: The more common attributes two concepts have, the bigger the interconnection-degree is. Different attributes should have different measures. The values of attributes affect the interconnection-degree. Whole-part interconnection is an important attribute-based one. Similar to Winston, Chaffin, and Hermann, we give a taxonomy about the part-whole relation in military domain. Moreover, there are two important questions in whole-part interconnection: one is whether parts inherit the values of attributes of the whole, and the other is the relations between the values of certain attributes of parts in cuts and one of the same attribute of the whole. P represents a tree consisting of all parts of certain whole. A cut is a subset of P,, and has exactly one common element with any the most offshoot of the tree. We can use whole-part interconnection to check inconsistency. For example, if A1 was cut-additional and C(A') ≠Σ_(C' ∈T)C'(A'), then the values of A1 are inconsistent. A' is called cut-additional if A' satisfies C(A')= Σ_(C' ∈T)C'(A') for any T of P. We described two interconnection algorithms in section 5. The purpose of the first interconnection algorithm about two concepts C_1 and C_2 is output as much knowledge as possible. The algorithm mainly has tree steps. The first step is to consider the relations interconnection between C_1 and C_2. The second is to consider attribute-based interconnection between C_1 and C_2. The third is to analyze attribute-values-based interconnection. The aim of second algorithm is to solve such problem as comparing attributes satisfying special probability distribution. Concept interconnections embody importance in the phrase of knowledge acquisition and knowledge service systems such as natural language query system.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.iscas.ac.cn/handle/311060/7572
Appears in Collections:中科院软件所

Files in This Item:
File Name/ File Size Content Type Version Access License
LW008645.pdf(2532KB)----限制开放-- 联系获取全文

Recommended Citation:
雷玉霞. 基于本体的军事知识联通[D]. 中国科学院软件研究所. 中国科学院软件研究所. 2002-01-01.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[雷玉霞]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[雷玉霞]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2017  中国科学院软件研究所 - Feedback
Powered by CSpace