首页 | 人工智能 | 数据挖掘知识 | 相关研究方向 | 编程技术 | 电脑常识 | 互联网资源 | 交流论坛 | 免费书籍资料下载 | 论文下载 | 文档资料 | 在线手册
人工智能: 信息检索 商业智能 搜索引擎技术与新闻 神经网络 生物信息学 模式识别 知识工程 本体理论与方法 机器学习 决策支持 自然语言理解 专家系统 >>更多
数据挖掘知识:
数据挖掘论文 数据挖掘其他 数据挖掘工具与应用 时序模式 相关研究人员主页 相关方向求职招聘信息 文本挖掘 学位论文 异类 预测 web数据挖掘 >>更多
相关研究方向: 联机分析 信息抽取 小波变换 数据仓库 access数据库 DB2数据库 Mysql数据库 Oracle数据库 SqlServer数据库 Sysbase数据库 统计分析 >>更多
主页>人工智能>本体理论与方法>

Is Ontology Overrated?

Today, I read the article: Ontology is Overrated : Categories, Links, and Tags finally, which has been stayed in my book list for more than one week.

字串6

"Is ontology overrated ?" It is a problem worthy of discussion. Currently more and more people is getting involved over this point. I find another related article just now: The Semantic Web, Syllogism, and Worldview (supposed to look into it later). Eventually, whatever the answer is we will  make things more clearly .  字串8

  字串8

1. Ontology vs. classification 字串8

Fisrt of all, I think it is a pity that although the aritcile is entitled  with "ontology is overratted" and that the ontology is mentioned and defined in the article, much more discussions are done around classification instead of ontology.

字串4

1.1. An ontology is not a classification schema

字串2

Yes, ontology and classification has something in common: they make some statements about the world(either the world as a whole a in a particular domain.)  But the statements are done in the totally different manners.

字串1

The purpose of ontology is to descrbie/explain concepts. while classifcation focused on classificatoin things. (Although sometimes concepts are organized terminologically, which makes people feel that ontology are similar to classification schema.) 字串7

With ontology, we try to explain what a stuff really is, not to build an judging system like classification does. So ontology in fact wants to struggle out of the difficult situations classification met with: no single perfect classificatoin system fit for the whole world. 字串7

We have different classification schemas based on different consideratoins(attributes of the realthings). Ontology provides those considerations we can mentioned about, based on which classification schema can be formed dynamically.

字串4

To take all kinds of attributes of real things into consideration is an essetial difference between ontology and classification schema.That makes ontology avoid the problem of single classification. 字串3

1.2. no perfect ontology for the whole world, too.

字串5

I have mentioned before, ontology is a conceptualization. Does there exists a perfect conceptualization for the whole world? Still the answer is NO. 字串1

So my point is that, the question "Is ontology overrated ?" is worthy of exploring, but the negative answer should not be concluded from the analysis on classification system. 字串6

2. ontology vs. tags

字串8

Most of the critics on ontology/semantic web comes from the tagging campaign. 字串2

Yes, I admit the usefulness and flexibility of tagging. I pay much attention into tagging too. But will tagging be an substitute (ender) of ontology? Of course NO. 字串7

what does tagging bring us?

字串1

We can′t find an organized web by tagging, which means tagging is not indexing, although we may mis-think it is. To my point of view, tagging is just like the common web pages we browsing everyday. It is information, it links to anywhere, it is semantically ambiguous.

字串7

The sparkle point of tagging is: people get evolved in. We can express(to ourselved, to computer, to the others in the world) what we think this article is about. By this kind of expression, 1)virtual communities can be found,and 2)machine need not have to do as much guessing (statistic) work as before.

字串4

But tags mess and we may get a tag-flood too. So tags should be managed. As search engine manages web pages.

字串5

One of many possible problems of tagging is the semantic ambiguous. e.g. How can we know MAC and APPLE has the same meaning? Or, we know they are the same ,but computers put them too far away for us to finding the other. 字串6

Someone argues that we just guess the distances based on the pages they tagged. This, just  proves what I′ve said that tags should be managed as the webpages.

字串9

But is the guessing-method the only way? Of course we can explain the semantics of tags by ontology.

字串5

3. human readable vs. machine readable

字串1

It is interesting that recently I begin to thinking that why semantic web get so many suspicion? Because semantic web put more attentions on the machines side instead of the side of human beging. (The target of semantic web is to build an machine readable web, so that we can provid better service to people. ) 字串3

While there are many we can do to provide a better service without machine understadable. Such as tagging and blogging. Tagging pay more attention to people directly. 字串3

They will meet somewhere. 字串8

4. Is ontology overrated?

字串5

Ontology makes a beautiful promise, how far can we go?  字串4

上一篇:The Semantic Search Engine   下一篇:Introduction to Text Indexing with Apache Jakarta Lucene
版权申明:本站信息收集自互联网,仅供学习参考使用。若有违法转摘您的作品请email我们及时删除!  
用户名: 新注册) 密码: 匿名评论 所有评论
评论内容:(不能超过250字,需审核后才会公布,请自觉遵守互联网相关政策法规。
Google
8 热门推荐
  • 本体库的建立与标注
  • 汉语诗的本体知识与语意检索
  • 軟體工程的知識本體
  • 顶层本体建模与中华文化
  • Bootstrapping Semantics on the Web:Meani
  • Position Paper: Ontology Construction fr
  • Towards the notion of gloss, and the ado
  • Bringing Communities to the Semantic Web
  • 基于本体的语言模型研究
  • Organization Ontology
  • 8 阅读排行
     
    版权所有:数据挖掘研究院 2004-2006 未经授权禁止复制或建立镜像
    增值电信业务经营许可证编号:皖B2-20040042 文网文:[2005]027号