找回密码
 欢迎注册
搜索
热搜: 活动 交友 discuz
查看: 3886|回复: 18

感觉开局还不错

[复制链接]
发表于 2010-1-11 03:59:28 | 显示全部楼层 |阅读模式
开了一个头,感到不错。希望能成一件事。
实验性本体已开始发表在BMKI专栏上:
http://www.miforum.net/bhf/english/yitong2/hypertension100110.owl
 楼主| 发表于 2010-1-15 04:50:20 | 显示全部楼层
Google查到以下有关本体研究的网站:

欧洲本体研究中心:
http://www.ecor.uni-saarland.de/home.html

日内瓦:
http://www.sim.hcuge.ch/ontology/

生物医学信息学:
http://bmir.stanford.edu/

(Buffalo Univ. 和Stanford  Univ.)国家本体研究中心:
http://ncor.buffalo.edu/

NLM医学本体研究:
http://lhncbc.nlm.nih.gov/lhc/se ... CMedicalOntology.vm

日本本体论坛:
http://abelard.flet.keio.ac.jp/ontology/
 楼主| 发表于 2010-1-25 12:09:25 | 显示全部楼层
下面传上我在国际医学术语学组织(IHTSDO)概念模型SIG上发表的帖子,因为论坛有严格的知识产权规定,我只能上传自己发表的意见:

An experimental ontology for hypertension
Hanfei Bao
Date: 12 Jan 2010
Dear colleagues,
An experimental ontology for hypertension is under exploration aimed at investigating medical knowledge integration, the operations of clinical information based on EHR, etc(see the attachment)
Comments and suggestions are welcomed.
Regards.
Hanfei Bao
 楼主| 发表于 2010-1-25 12:10:21 | 显示全部楼层
Hanfei Bao 15 Jan
Thank Cathy and Pat.
Except using Protégé 4, one can open the file using programs such PSPad, even I.E. as XML-like format.
Pat’s suggestions are of great value. They are the most things to do for an ontology development.
The “hasFacilitating” (perhaps using “canFacilitate” is better) is taken here as one of so called meta-properties meaning a kind of active binary relations such as “increase” (in number, power, speed), “activate”, “strengthen”, etc.
The hierarchy structure of the ontology will be reconsidered repeatedly for more consistency. Because usually the axis systems for classification are dependent of each other in their definitions but complicatedly jointed in their extensions, to build the hierarchies of single axis system or less axis systems is always challenging.
I wish to develop a disease-oriented ontology as an open source to support mining of clinical data or e-health references and for medical knowledge engineering researches as well, through the way as following:
Machine-readable clinical data based on standard terminology and models <- communicating applications + computing applications → Machine-readable medical knowledge and machine-readable typical facts (Ontologies).
This version attached is optimized by an intrinsic function of Protégé 4.(see the attachment).
Hanfei Bao
 楼主| 发表于 2010-1-25 12:10:42 | 显示全部楼层
Hanfei Bao 17, Jan
I tried to define Pulse Pressure as following:
PulsePressure ≡ isDifferenceBetween(SystolicBloodPressure, DiastolicBloodPressure)
But It seems that Protege OWL-DL does not support this form. I can not help to do like this:
1. create object properties hasOperand-hasMinuend,hasSubtrahend;
2. select class PulsePressure and click the superclass(+) in its description view;
3. type in the text “hasMinuend some SystolicBloodPressure and hasSubtrahend some DiastolicBloodPressure” to finish the definition.
I feel no good about this method. I would ask some colleague on this forum if there are some better ways to do it.
Hanfei Bao
 楼主| 发表于 2010-1-30 17:29:51 | 显示全部楼层
新版本信息:

The version 100130 of the experimental ontology for hypertension involves the questions of the independent or related dimensions, existence or nonexistence, unit of a measurement, conceptual symptom and descriptive symptom, etc.

see: http://www.miforum.net/bhf/english/yitong2/hypertension100130.owl
 楼主| 发表于 2010-2-25 12:12:38 | 显示全部楼层
新版本信息,欢迎积极参与,袖手旁观可不符合学者的精神风貌

The new version of hypertension20100230.owl
The new version of hypermechanism20100230.owl
The new version of hypermedication20100230

地址为:  http://www.miforum.net/bhf/english/yitong2/calresource.html
 楼主| 发表于 2010-3-16 09:46:14 | 显示全部楼层
计算本体的胚型已发表:

http://www.miforum.net/bhf/engli ... mputing20100330.owl

正在加紧开发第一个application以对电子病历进行知识支持.
 楼主| 发表于 2010-3-27 16:31:53 | 显示全部楼层
1,计算本体修改:
http://www.miforum.net/bhf/english/yitong2/hypercomputing20100430.owl

2,干涉XML及OWL文件的java初级程序已实现运行
发表于 2010-3-28 01:30:28 | 显示全部楼层
包老师,
isDifferentFrom, owl 本身就有关键词,叫做owl:disjointWith

http://www.w3.org/TR/2002/WD-owl-ref-20021112/

血压等模型,可以看看 http://www.co-ode.org/galen/

推理,取决于您最后的目的,

Protege 支持SWRL http://www.w3.org/Submission/SWRL/?target=_blank

预祝您项目顺利
 楼主| 发表于 2010-3-28 10:02:56 | 显示全部楼层
谢谢猿人提醒。
这一项目希望与临床数据处理支持密切结合,也是我半生学术追求(生物医学知识工程)的最终目的。在此起始阶段,受到一些优秀学术青年的大力支持。这次在上海分别与石头和丹尼两位论坛领导作了深入探讨,准备先成立一个兴趣小组,欢迎优秀的,富有献身和开创精神的学术青年一起参加,由我先起草一个发展思路和策略供讨论。

我还有一个问题,Protege似乎非常刻板(也许是全球web背景共享的需要)。如你用Protege打开我们的Hypercomputing20100430.owl, (经修改后)你可以发现AbnormalSingleNumericalValueOfAttribute的父类为SingleNumericalValueOfAttribute,但我发现这是Protege的“表面文章”,在真正的XML文件(可用PSPad打开)中,AbnormalSingleNumericalValueOfAttribute的父类却始终是原来的名称,即SingleNumberOfValueOfAttribute。似乎错误的东西无法改正,这好像有点“淘浆糊”。
发表于 2010-3-29 02:00:34 | 显示全部楼层
包老师,我们做开发,都是用topbraid composer,您可以看看。protege有些超前尝试新技术,所以有时候不稳定。
 楼主| 发表于 2010-3-29 09:11:29 | 显示全部楼层
本帖最后由 老包 于 2010-3-29 09:12 编辑

谢谢猿人!

看来Protege问题还不小,我已经是第二次上当。第一次是根据Protege的介绍设置了Preferences,结果是开发的owl文件的元素名称均为随机码,只有Protege能读得懂,开发者自己读不懂,必须经过复杂的转化才能读懂自己的东西。

Protege现在似乎很有点独大的味道。我觉得这对服务于全球的开源项目是个大忌。
发表于 2010-3-29 12:56:26 | 显示全部楼层
谢谢猿人提醒。
这一项目希望与临床数据处理支持密切结合,也是我半生学术追求(生物医学知识工程)的最终目的。在此起始阶段,受到一些优秀学术青年的大力支持。这次在上海分别与石头和丹尼两位论坛领导作了深入探 ...
老包 发表于 2010-3-28 10:02


这个提议很好.
 楼主| 发表于 2010-4-2 09:13:02 | 显示全部楼层
本帖最后由 老包 于 2010-4-2 09:18 编辑

1.小组成员完成标准自然陈述语句结构的RDF式自动解析,见:
http://www.miforum.net/bhf/english/yitong2/ClassOfCKEFToRDF.doc
2. 对自然语言陈述句的主体成分和客体成分的结构分析(一般序修饰和关联词修饰分析)又形成新的设计思路。
您需要登录后才可以回帖 登录 | 欢迎注册

本版积分规则

快速回复 返回顶部 返回列表