外文翻译-在人才招聘流程中的语义Web技术的影响

1、在人才招聘流程中的语义Web技术的影响摘要：在网上招聘领域中使用语义Web技术，可以大大增加市场的透明度，为雇主降低交易成本，改变中介机构所涉及的商业模式。在本文中，我们描述了如何网上招聘流程能够简化使用语义Web技术。我们分析了这样一个从技术和商业角度情况下的方案的实现机会。我们描述了一个典型的实现所需的技术基础设施和分析潜在的经济影响，有关当事人的利益和他妈的商业模式的潜在变化。关键词：语义Web，招聘流程，电子市场，本体1 引言互联网已近发展成招聘和就业过程中的主要媒介。最近在德国填补的职位超过53%的是网上发布工作的结果，并根据大多数预测这一比例有可能进一步增加WiMo03。尽管取得了

2、这些成就，在网上劳动力市场的信息流是远离理想的。大量的网上求职门户如雨后春笋般出现，分成网上劳动力市场信息岛屿，使求职者接近不可能得到所有有关未平仓合约的概述。他们的强大的市场地位作为强大的出发点，为求职者允许工作门户，以高昂的费用收取雇主出版平仓。因为这些费用雇主发布他们的招聘职位，仅是一小部分的门户网站，从而达到所有合格的申请人的报价。由于劳动力市场的紧张局势，雇主经常收到大量的申请开仓。手动预选候选人的成本上升和雇主对寻找自动预选候选人的重要性。因此，雇主想向预选合适的人选来发布招聘信息，以减少交易成本。求职者将得益于增大的透明度，从劳动力市场寻找一个匹配的职位。在本文中，我们认为这两个

3、目标可以同时达到通过在招聘流程中使用语义Web技术。语义WebBeHe01的基本思想是扩展Web，机器可理解的数据，除了传统的HTML页面。这个目标是利用网络就像一个全球式的分布式的知识存储，能被杠杆式投机通过应用程序来自动执行任务。为实现这一目标的基本技术是：统一资源标识符（URL）来作为一个全球资源和术语来描述他们的识别机制,资源描述框架（RDF）作为基本的数据模型连同基于XML的序列化语法一起发布在网络上的数据Fehler! Verweisquelle-编著gefunden werden。,Web本体语言（OWL），表现知识表示的条款和概念扩展RDFW3C04c。语义Web的发展是科学（

4、麻省理工学院，斯坦福大学ILRT等）和万维网联盟（W3C）带领的商业机构（惠普，IBM，诺基亚等）的加入的成就。有大量的研究和语义技术的发展投资的IT市场QuKa04的市长参与者。已经部署的在生物医学和书目数据整合Schr04，QuKa04领域的第一次大规模的原型。语义Web的技术基础日趋成熟。目前还不清楚新技术对市场和企业的经济效果。与我们的研究，我们的方法从业务语义技术的潜在影响和技术的角度看，为了使有关的新技术对市场，企业和个人的影响的预测。谁将受益？谁就会失去自己的市场地位，或将不得不改变自己的商业模式？这与分析技术限制相结合的参与者的角色分析，使我们的项目是否用一个方案是可能实现的。

5、在本文的第2部分，我们描述了如何使用语义Web技术使网上招聘流程能够被简化。我们描述了在第3节中的原型实施所需的技术，侧重于使用本体的数据集成基础设施和语义匹配服务。在第4节中，我们从不同参与者的观点情况下实现的潜在经济影响进行分析。2 语义Web技术支持的招聘流程从组织的角度来看，一个通常的招聘过程分为四个阶段：需求分析，发布招聘启示，接收和预先选择的应用程序，最终招聘的决定。在本节中，我们描述了如何在不同的阶段可以通过语义Web技术的支持。中心构建模块的情况下使用受控词表来注释招聘信息和应用，并纳入包括这些描述的流程文件来使用RDF。图1给出了下面过程描述的概述。图1：使用语义Web技术的

6、招聘流程2.1 需求分析未平仓合约的要求通常由经营部门与人力资源部门一起构思。如今招聘启事描述的结果，对公开职位的要求是写在自由文本上的。自由文本的使用限制在招聘过程后期阶段的记录的计算机的可加工性。相反免费的文字说明，使用一个共同的语言以受控词表的形式来描述将发布在流程里面的自动化潜力的各项任务的招聘信息的细节。这个额外的努力编写语义注释的缺点可能会超过通过提高效率的储蓄。2.2 发布招聘启事有大量的商业工作门户，像Monster，JobPilot或StepSone，争相为出版费张贴发布工作。不同的门户代表不同的观众，并在某一地理区域专业化。由于出版费用容易合计，公司只有数量有限的门户发布自

7、己的帖子。张贴第二次发布的渠道是公司自己的网站。企业网站上发布的帖子是便宜的，但是只达到的观众是非常有限的，因为目前的搜索引擎像谷歌的索引功能来支持平仓的收缩是太不精确的。为了提高市场透明度一些公共机构,如德国联邦就业办公室(BA)和瑞典国家劳动力市场管理局(AMS)已经开始的项目将空缺职位在一个中央数据库。在两个项目中,参与雇主使用条款从控制vocabularyto分类他们的帖子,送他们到中央数据库使用hr - xml数据格式的变化。收集到的帖子发表通过一个中央门户和转发到私营工作门户网站。这些项目是整个市场的问题取决于一个中央数据库一个许多mrket参与者对象和方法可以非常容易出错xper

8、ience德国项目显示Cross03。使用语义Web技术销售职位将增加市场透明度一起避免中央数据库的瓶颈。组织将发布注释工作职位直接在他们的网站使用条款从一组受控词汇表和RDF数据格式。然后帖子可以直接爬所有感兴趣工作门户网站,这意味着集中的方法将被替换为一个分布式的方法。因此,所有的工作门户操作相同的信息,帖子将达到更多的申请者,导致更高的市场透明度。另一个好处从帖子从受控词汇表术语的注释是条款可以结合一个工业领域的背景知识。工作门户网站可以提供语义匹配服务从而计算工作职位和申请人之间的语义相似度,基于背景知识概要文件如何不同的相关条款。例如,如果Java编程技能是需要一定的工作,申请人在D

9、elphi经验丰富,匹配算法会考虑这个人的形象比别人的更好的匹配熟练地在SQL,因为德尔菲比SQL和Java和Java是密切相关的。这种方法允许职位信息的比较和申请人的档案利用背景知识,而不是仅仅依靠控制关键词,像传统的搜索引擎。2.3工作申请的接收和预选雇主经常收到大量的应用程序为一个开放的位置,由于劳动力市场的紧张状况。为了降低处理成本的雇主已经开始喜欢在经典论文在线应用程序的应用程序。其他公司如IBM或大众建立了web表单接收应用程序,申请人必须重新输入他们的应用程序的基本事实的机器可处理的方法。这个数据是用来自动候选人预选满足基本要求,从而减少了预选的成本。申请人和雇主都将受益的语义注

10、释工作申请除了古典自由文本应用程序。应用程序可以重用RDF概要文件并寄给不同的雇主,而不是填充数据到不同的web表单。雇主使用自动化的注解预选过程匹配的最小需求的应用程序。这个阶段的另一个问题是防奖状和教育的经验。这些证据可以附加到应用程序的形式进行数字签名的RDF语句发行大学或组织。打样的过程中数字签名推荐可以自动化,进一步降低成本。2.4面试和招聘的决定面对面的接触为最后的招聘决定候选人是必不可少的。因此,招聘过程的这个阶段,候选人面试,他们的柔软和技术技能评估通过面对面的会议,甚至在评估中心在未来仍将是手动完成的。3技术实现在本节中,我们提出我们的原型实现所需的技术基础设施实现上面描述的

11、场景。我们专注于如何来自标准词汇表已经在使用域内,对数据集成基础设施上可以加上现有non-RDF人力资源系统和匹配算法计算相似性的工作要求和申请人的资料。我们典型的实现是基于现有的开源组件,我们扩展场景特定功能。我们的劳动力市场场景暗示三个基本角色:信息提供者发布空缺职位和公司背景infoemation使用受控词汇表和RDF数据格式。信息聚合器谁爬发布信息并将其呈现给用户所需的格式和聚合。结合招聘信息和背景知识对工业部门允许聚合器提供语义匹配服务。使用url雇主识别机制允许帖子的浓缩与雇主的信息从不同的外部信息服务。信息门户的用户使用一个选择中央发布的所有访问点职位和背景有关雇主的信息,而不是

12、从不同的网站收集信息片段。图2说明了场景和角色模型技术架构的概述。图2:场景架构和参与者的角色3.1本体雇主和工作门户网站之间的数据交换是基于一组词汇提供共同的术语来描述工作类型,所需的技能和教育背景。广泛的标准过程文件和分类模式已经为招聘开发。在下面,我们概述了相关标准和描述我们如何整合他们的本体所使用的原型。3.1.1文档处理过程使用最广泛的标准过程文件像招聘信息和应用程序是hr - xml标准。hr - XML HR-Consortium已经同意,一个独立的非盈利组织,致力于开发和推广标准的XML规范,使电子商务套件和人类resources-related数据交换的自动化。hr - XM

13、L是一个图书馆的75多个相互依存的XML模式。XML模式定义的数据元素特定的人力资源事务,以及选择和约束管理这些元素的使用。hr - xml财团已经产生了计划覆盖的主要流程,以及组件方案,使用跨多个业务流程。3.1.2分类法这些是几个分类法用于技能的描述,分类工作和公司所属的工业部门。他们提供的词汇中使用过程文档。标准职业分类系统(SOC)将员工分为职业类别(23个主要群体,96小群体,449职业)。每个职业都包含详细的职业(s)要求类似的工作职责,技能,教育和经验。职业代码(Berufskennziffer - BKZ)是德国版的SOC系统。它将员工分为5597职业类别根据职业定义。行业分类

14、(WZ2003)是一个标准的经济sctivities分类排列,这样可以将实体根据特定的活动。北美行业分类系统(NAICS)提供了行业sectof定义为加拿大、墨西哥和美国促进经济分析。3.1.3我们人力资源本体在第一步中实现我们的人力资源的情况下,我们已经创建了一个人力资源本体(人力资源本体)通过集成现有的广泛使用的标准和分类。为了避免冗余我们sub-ontologies形式(见图3)用于工作应用程序的描述。本体的基本结构主要来源于hr - xml标准。我们适应了工作位置导引头和职位发布部分从HR-BA-XML(德国版的hr - xml)。指定行业sub-ontology我们利用两个德国工业部

15、门的分类(WZ2003)和北美行业分类系统(NAICS)。作为职业分类,我们利用德国占领代码(BKZ)和标准职业分类系统(SOC)。图3:在HR-ontology Sub-ontologies技能sub-ontology我们定义概念代表能力,这是用来描述工作要求以及申请人的技能。这个本体还提供了对指定某一技能的能力水平。我们得到的技能sub-ontology KOWIEN本体(Alan03)。此外,我们开发了一个人sub-ontology描述求职者的个人信息和组织sub-ontology指定雇主的公司。我们的人力资源本体建模使用OWL(W3C04c)为了让精确的模式验证输入数据。工作职位的描述

16、和申请人的概要文件存储在RDF(W3C04a)使用由人力资源本体定义的词汇。3.2数据集成基础设施雇主可以使用两种不同的方法来发布注释工作职位根据自己现有的软件基础设施。如果使用数据库支持标准软件,它们可以使用映射关系数据直接导出到RDF工具如Bize03,Bize04。如果他们不使用任何工具来管理职位,他们可以注释HTML版本的帖子使用注释工具像KaHe02。创建RDF数据发表使用标准的Web服务器,比如Apache或专门的查询接口的MoSe03。工作门户呈现的求职者将公布的数据:使用一个目录来定位信息提供者;使用爬虫收集数据从不同的供应商;使用映射引擎将公布的数据使用不同的词汇;提供语义匹

17、配服务匹配申请人的简介和招聘信息基础上的知识包括本体和可用的市场数据;呈现HTML数据根据用户的偏好(详细级别,设备,语言等等)。上面的架构后,门户将RDF存储大量的收集数据并提供高性能访问这些数据给他的客户。专业数据库和查询语言的概述中可以找到Beck03。在我们的原型使用D2RBize03映射关系数据映射到RDF,NetAPI的MoSe03和说唱(OlBi04)发布数据和Jena2框架Carr04扩展与语义匹配引擎门户实现的基础。4经济影响提高市场透明度的可能性提出了新的战略问题的市场参与者将最终决定他们愿意参与:场景在经济激励各方参与?本节将简要讨论的可能影响我们的场景的实现商业模式的参

18、与者。网上招聘信息门户是求职者和雇主之间的中介。收取发表费,门户网站的业务模式正在使用今天阻止雇主发布他们的帖子在更大数量的工作门户网站。这导致一个部门在线劳动力市场信息的岛屿和使它接近一个部门在线劳动力市场信息的岛屿和使它几乎不可能对一个求职者的概述所有相关职位。使用语义Web技术将改变信息流。开放的位置将使用雇主的网站上发布的所有工作门户网站。原通信链接控制改变开放的n:m沟通情况。这将允许求职者使用一个门户的访问所有可用信息的市场。雇主将节省出版费用通过发布职位只有在他们自己的网站上。注释的额外成本工作职位可能储蓄效应的出版费用,通常几百欧元。关键问题是就业门户网站的商业模式在语义Web

19、的场景。门户如何融资服务,当他们不能依靠出版费用吗?有几个选项。可以由门户网站服务费支付的求职者获取市场数据和额外的服务,如语义匹配;财团的雇主,形式以达到更多潜在的申请人和节省出版费;政府当局有兴趣平滑信息流动的劳动力市场以降低失业率。另一个问题是融资维持在市场上使用的受控词汇表。这也可以是一个公共机构的任务比如德国联邦就业办公室(BA)作为一个市场或市场参与者可以通过一个非营利性的财团,类似于在hr - xml财团已经发生的过程。5结论和展望在本文中,我们提出了一个场景和语义Web技术支持招聘流程。我们描述了在线招聘过程,分析了在这一领域最先进的技术。我们建议改进的几个步骤提高职位和工作应

20、用程序与注释使用受控词汇表。我们描述的原型实现的场景。原型是一个人力资源的核心组件本体来自现有规范,数据发布和集成基础设施和检索引擎使用语义相亲。本体的创建和原型实现的软件基础设施显示我们的场景在技术上可以实现的。分析参与者的利益和潜在的经济影响表明就业市场的主要演员-雇主和求职者都受益于场景的实现。求职者将受益于增加市场透明度。雇主将受益于达到更多潜在的申请者,省钱的出版工作职位和节省成本的预选自动化应用程序。我们未来的工作将集中在完成原型,建立一个更大的带注释的工作职位的实际数据集测试基础设施的可伸缩性,并提高语义匹配算法。The Impact of Semantic Web Techno

21、logies on Job Recruitment ProcessesAbstract: Using Semantic Web Technologies in the domain of online recruitment could substantially increase market transparency, lower the transaction costs for employers, and change the business models of the intermediaries involved. In this paper, we describe how

22、online recruitment processes can be streamlined using Semantic Web technologies. We analyze the realization chances of such a scenario from the technological and business viewpoint. We describe a prototypical implementation of the required technological infrastructure and analyze the potential econo

23、mic impacts, the interests of involved parties and potential changes to their business models.Keyword: Semantic Web, Recruitment Process, Electronic Market, Ontology1 IntroductionThe Internet has already evolved into primary medium for recruitment and employment processes. Over 53% of recently fille

24、d positions in Germany are the result of an online job posting and according to most forecasts this share is likely to increase further WiMo03.Despite these achievements, the information flow in the online labour market is far from optimal. A large number of online job portals have sprung up, dividi

25、ng the online labour market into information islands and making it close to impossible for a job seeker to get an overview of all relevant open positions. Their strong market position as the prime starting point for job seekers allows job portal to charge employers high fees for publishing open posi

26、tions. Due to these costs employers publish their job postings only on a small number of portals, which prevents the offers from reaching all qualified applicants.Employers often receive a large number of applications for an open position, due to the strained situation of the labour market. The cost

27、s of manually pre selecting potential candidates have risen and employers are searching for means to automate the preselection of candidates.Thus, employers would like to decrease transaction costs for publishing job postings and for preselecting appropriate candidates. Job seekers would profit from

28、 increased transparency in the labour marker in their search for a matching position. In this paper, we argue that both goals could simultaneously be reached by using Semantic Web technologies within recruitment processes.The basic idea of the Semantic Web BeHe01 is to extend the Web, in addition to

29、 classic HTML pages, with machine-understandable data. The objective is to use the Web like a global distributed knowledge store which can be leveraged by application to perform tasks automatically. The base technologies for realizing this vision are:Uniform Resource Identifiers (URIs) as a global i

30、dentification mechanism for resources and terms used to describe them,The Resource Description Framework (RDF) as basic data model together with its XML-based serialization syntax for publishing data on the Web Fehler! Verweisquelle konnte nicht gefunden werden.,The Web Ontology Language (OWL) exten

31、ding RDF with terms and concepts for expressive knowledge representation W3C04c.The development of the Semantic Web is a joined effort of scientific (MIT, Stanford, ILRT etc.) and business institutions (HP, IBM, Nokia etc.) led by the World Wide Web Consortium (W3C). There are substantial research a

32、nd development investments in semantic technologies by the mayor players of the IT market QuKa04. The first large scale prototypes have been deployed in the domains of biomedical and bibliographic data integration Schr04, QuKa04.The technological foundations of the Semantic Web are getting mature. W

33、hat is still unclear are the economic effects the new technology will have on markets and enterprises. With our research, we approach potential impacts of semantic technologies from the business and the technological viewpoint, in order to make predictions about the influence of the new technologies

34、 on market, enterprises, and individuals. Who will benefit? Who will lose his market position or will have to change his business model? This analysis of the participants roles combined with the analysis of technological restrictions allows us to project whether a scenario is likely to be realized.I

35、n section 2 of this paper, we describe how online recruitment processes can be streamlined using Semantic Web technologies. We describe the prototypical implementation of the required technological in section 3, focusing on the ontologies used, the data integration infrastructure, and on semantic ma

36、tching services. In section 4, we analyze the potential economic impacts of the realization of the scenario from the perspectives of the different participants.2 Supporting Recruitment Processes with Semantic Web TechnologiesA usual recruitment process seen from the organizations viewpoint can be di

37、vided into four phases:1. Requirements analysis,2. Publishing the job posting,3. Receiving and pre-selecting applications and4. Final recruitment decision.In this section, we describe how the different phases can be supported by Semantic Web technologies. The central building blocks of the scenario

38、are the use of controlled vocabularies to annotate job posting and applications, and the use of RDF to include these descriptions into process documents. Figure 1 gives an overview of the process described below.Figure 1: Recruitment process using Semantic Web technologies.2.1 Requirements AnalysisT

39、he requirements for an open position are usually conceived by the operating department together with the human resource department. Nowadays the resulting job posting describing the requirements for an open position is written in free text. The usage of free text limits the machine processability of

40、 the postings in the later phases of the recruitment process. In contrast to free text description, the usage of a common “language” in the form of a set of controlled vocabularies to describe the details of a job posting would open up the potential of the automation of various tasks within the proc

41、ess. The disadvantage of the additional effort for writing semantic annotations could be outweighed by savings through increased efficiency.2.2 Publishing the job postingThere are a large number of commercial job portals, likenMonster, JobPilot or StepSone, competing to publish job posting for a pub

42、lication fee. The different portals represent various audiences and are specialized in certain geographic regions. As publication fees easily add up, companies publish their postings only on a limited number of portal. The second publication channel for job postings is the companys own website. Publ

43、ishing postings on the corporate website is cheap but reaches only a very limited audience, because the indexing capabilities of current search engines like Google are too imprecise to support directed searches for open positions.In order to improve market transparency several public bodies like the

44、 German Federal Employment Office (BA) and the Swedish National Labour Market Administration (AMS) have started projects to integrate open positions in a central database. In both projects, participating employers use terms from a controlled vocabularyto categorize their postings and send them to th

45、e central database using variations of the HR-XML data format. The collected postings are published through a central portal and are forwarded to privately run job portals. The problem with these projects is that the whole market depends on one central database an approach to which many mrket partic

46、ipants object and which can be very error-prone as xperience from the German project shows Cross03.Using Semantic Web Technologies to distribute job posting would increase market transparency together with avoiding the bottleneck of the central database. Organizations would publish annotated job pos

47、tings directly on their websites using terms from a set of controlled vocabularies and the RDF data format. The postings could be then crawled directly by all interested job portals, meaning that the centralized approach would be replaced with a distributed approach. Consequently, all job portals wo

48、uld operate on the same information and postings would reach more applicants, resulting in higher market transparency.Another benefit from having postings annotated with terms from a controlled vocabulary is that the terms can be combined with background knowledge of an industrial domain. Job portal

49、s could offer semantic matching services which would calculate the semantic similarity between job postings and applicants, profiles based on background knowledge about how different terms are related. For example, if Java programming skills are required for a certain job and an applicant is experie

50、nced in Delphi, the matching algorithm would consider this persons profile a better match than someone elses who is skilful in SQL, because Delphi and Java are close related than SQL and Java. This approach allows the comparison of job position postings and applicants profiles using background knowl

51、edge instead of merely relying on the containment of keywords, like traditional search engines do.2.3 Receiving and preselecting job applicationsEmployers often receive a large number of applications for an open position, due to the strained situation of the labour market. In order to reduce handlin

52、g costs employers have begun to prefer online applications over classic paper applications. Others companies like IBM or Volkswagen have set up web forms for receiving applications, where applicants have to re-enter the basic facts of their application in a machine processable way. This data is used

53、 for the automated preselection of candidates fulfilling the basic requirements and thus reducing the preselection costs.Both applicants and employers would benefit from the semantic annotation of job applications in addition to the classical free text application. Application could reuse their RDF

54、profile and send them to different employers rather than fill the data into different web forms. Employers use the annotations for automating the preselection process by matching their minimal requirements with the applications.Another issue of this phase is proofing testimonials of experience and e

55、ducation. These proofs could be attached to applications in the form of RDF statements which are digitally signed by the issuing university or organization. The process of proofing digitally signed testimonials can be automated, further reducing costs.2.4 Interviews and Recruitment DecisionFace-to-f

56、ace contact with candidates is indispensable for the final recruitment decision. Thus, this phase of the recruitment process, where candidates are interviewed and where their soft and technical skills are evaluated through face-to-face meeting or even in assessment centres will still be done manuall

57、y in the future.3 Technical RealizationIn this section we present our prototypical implementation of the technical infrastructure needed to realize the scenario described above. We focus on how vocabularies can be derived from standards already in use within the domain; on how the data integration i

58、nfrastructure can be coupled with existing non-RDF human-resource systems and on matching algorithms for calculating the similarity of job requirements and applicants profiles. Our prototypical implementation is based on existing open source components which we extend with scenario-specific features

59、.Our labour market scenario implies three basic roles:Information Providers which publish open positions and company background infoemation using controlled vocabularies and the RDF data format.Information Aggregators who crawl the published information and present it to the information users in the

60、 desired format and aggregation. Combining job postings with background knowledge about the industrial sector allows the aggregators to offer semantic matching services. The usage of URLs as employer identification mechanism allows the enrichment of postings with information about employers from dif