李健興 長榮大學資訊管理系副教授 兼資訊工程系籌備處主任 2004/11/17

Presentation on theme: "李健興 長榮大學資訊管理系副教授 兼資訊工程系籌備處主任 2004/11/17"— Presentation transcript:

1 李健興 長榮大學資訊管理系副教授 兼資訊工程系籌備處主任 2004/11/17
建構CMMI知識地圖 李健興 長榮大學資訊管理系副教授 兼資訊工程系籌備處主任 2004/11/17

2 Outline Introduction The Structure of Ontology
Ontology-based Knowledge Management System Ontology Construction CMMI Ontology CMMI Assistant Tools CMMI Ontology Extraction Future Works

3 Introduction

4 Ontology (知識地圖) The ontology is a collection of key concepts and their interrelationships collectively providing an abstract view of an application domain. An ontology is a formal, explicit specification of a shared conceptualization. Conceptualization Explicit Formal

5 Ontology (知識地圖) Ontology–explicit formal specifications of the terms in the domain and relations among them. An ontology contains a hierarchy of concepts within a domain and describes each concept’s property through an attribute-value mechanism. Relations between concepts describe additional logical sentence.

6 Ontology (知識地圖) The main application areas of ontology technology
Knowledge management Web commerce Electronic business Database design Natural language processing Multi agent system ………

7 Example 交通資訊 台北台南 台北高雄 … 台北澎湖 火車 飛機 自行開車 搭巴士 搭船 班次: 車種: 時間: 速度: 價格:
航空公司: 班機號碼: 時間: 速度: 價格: 路線: 時間: 巴士公司: 路線: 時間: 價格: 船公司: 路線: 時間: 價格:

8 Ontology Example 氣象 Relation Association 氣象報導 氣象百科 天文 . . . . . . 寒流
颱風 降雨 中央氣象局/氣象局 型態：預報人員、 天氣圖 表示、警告、評估 颱風 編號:***(Neu)號 中心位置：:***(Nc)(Ncd)(Neu)(Nf) 強度:輕度颱風 型態:暴風圈 來襲、形成、登陸 氣壓 型態:副熱帶高氣 壓、熱帶性 低氣壓 增強為、逼近 降雨 降雨量***(Neu)公釐 累積雨量***(Neu)公釐 種類:大雨、陣雨、 大雷雨、豪雨 、豪大雨 型態：雨量、打雷 發生、襲擊、增加 災害 型態:水災、旱象、 土石流、山崩 、洪水、房屋 倒塌、河水暴 漲、落石、雷 擊、霜害 來襲、形成、登陸 造成 導致、造成、帶來 發佈、 表示 提醒 這是經由FIA推論後得到的event ontology。 發生 民眾/人民 型態:人數 注意、受困 向、往 導致 帶來、引進 地區 區域:山區、平地、 台灣、中部、 東半部 各縣市：台北市、台 南縣 海域:東海、南海 海岸:西海岸、沙岸 呈現、滯留、徘徊 影響 氣流 型態:西南氣流、 冷氣流 接近、影響、流動 時間 型態:最近、昨日、 今日、白天、 午後 根據、開始 移動方向 方向:東方、南方 西北方、東 南方 移動、靠近、前進 農林漁牧業 型態:漁港、農田 、農作物、 魚貨量 避風、休耕

9 DAML+OIL format <?xml version=‘1.0’ encoding=‘Big5’?> <rdf:RDF xmlns:rdf = xmlns:rdfs= xmlns:daml=” xmlns:xsd = xmlns:a =” > <daml:Ontology rdf:about=”氣象”> <daml:imports rdf:resource=” /> </daml:Ontology> <daml:Class rdf:ID=”氣象”> </daml:Class> <daml:Class rdf:ID=”氣象報導”> …… …… …… <daml:range rdf:resource=” #災害”/> </daml:ObjectProperty> <daml:ObjectProperty rdf:ID=”影響”> <daml:domain rdf:resource=” #災害”/> <daml:range rdf:resource=” #農林漁牧業”/> </daml:ObjectProperty> </rdf:RDF> 再將event Ontology轉成電腦可理解的XML format。

10 Characteristics of Ontology
Formal Semantics Consensus of terms Machine readable and processable Model of real world Domain specific

11 Reasons to Develop Ontologies
To share common understanding of the structure of information among people or software agents. To enable reuse of domain knowledge. To make domain assumptions explicit. To separate domain knowledge from the operational knowledge. To analyze domain knowledge.

12 Process of Developing an Ontology
Developing an ontology includes: Determine the domain and scope of the ontology. Consider reusing existing ontologies. Enumerate important terms in the ontology. Define classes in the ontology and arrange the classes in a taxonomic (subclass-superclass) hierarchy. Define attribute and describe allowed values for these attribute. Fill in the values for attribute for instance.

13 Ontology Learning Process

14 The Structure of Ontology

15 The three-layered object-oriented ontology
Domain Association Generalization Aggregation … Category 1 Category 2 Category 3 Category k Concept 1 Attributes 1 Operations 1 Concept 2 Attributes 2 Operations 2 Concept 3 Attributes 3 Operations 3 Concept n Attributes n Operations n … Concept 4 Attributes 4 Operations 4 Concept 5 Attributes 5 Operations 5 Concept 6 Attributes 6 Operations 6 … … Concepts Set

16 The four-layered Object-Oriented Ontology
Domain Category 1 Category 2 Category k Concept 3 Attributes 3 Operations 3 Concept 1 Attributes 1 Operations 1 Concept 2 Attributes 2 Operations 2 Concept n Attributes n Operations n Concept 4 Attributes 4 Operations 4 Class-layer Instance 3 Instance 1 Instance 2 Instance m Attributes m Operations m Instance 4 Instance-layer Association Generalization Aggregation Instance-of

17 The four-layered News Ontology (cont.)
Domain Event 1 Event k Event 3 Event 2 … Candidate Chinese Terms ……… Extended Concept Relation Category 1 Category 2 Category q Association Operation Templates Attributes Event Concept 1 Object Concept 1, m Event Concept 1, 1 Event Concept P Event Concept P,1 Object Concept P,n

18 The four-layered News Ontology
氣象 Relation Association 氣象報導 氣象百科 天文 寒流 颱風 降雨 中央氣象局/氣象局 颱風 氣壓 降雨 災害 編號:***(Neu)號 中心位置：:***(Nc)(Ncd)(Neu)(Nf) 強度:輕度颱風 型態:暴風圈 型態:副熱帶高氣 壓、熱帶性 低氣壓 降雨量***(Neu)公釐 累積雨量***(Neu)公釐 種類:大雨、陣雨、 大雷雨、豪雨 、豪大雨 型態：雨量、打雷 型態:水災、旱象、 土石流、山崩 、洪水、房屋 倒塌、河水暴 漲、落石、雷 擊、霜害 型態：預報人員、 天氣圖 表示、警告、評估 增強為、逼近 導致、造成、帶來 造成 發佈、 表示 提醒 來襲、形成、登陸 來襲、形成、登陸 發生、襲擊、增加 民眾/人民 帶來、 引進 發生 向、往 遠離、移動 導致 型態:人數 地區 恢復 注意、受困 影響 區域:山區、平地、 台灣、中部、 東半部 各縣市：台北市、台 南縣 海域:東海、南海 海岸:西海岸、沙岸 出現、發生 氣流 移動方向 時間 農林漁牧業 型態:西南氣流、 冷氣流 方向:東方、南方 西北方、東 南方 型態:漁港、農田 、農作物、 魚貨量 型態:最近、昨日 今日、白天 午後 影響 接近、影響、流動 移動、靠近、前進 避風、休耕 根據、開始 呈現、滯留、徘徊

19 Fuzzy Ontology (cont.) Generalization Aggregation Domain Association
Category 1 Category 2 Category 3 …………… Category k LBR Event E1 Event E2 Event E3 …… Event Ep LNR C : Concept A : Attribute O : Operation {C1;μC1E1,μC1E2,…,μC1Ep} {C2;μC2E1,μC2E2,…,μC2Ep} {C3;μC3E1,μC3E2,…,μC3Ep} AC11,AC12 ,…,AC1q1 AC21,AC22 ,…,AC2q2 AC31,AC32 ,…,AC3q3 OC11 ,OC11,…,OC1q1 OC21 ,OC21 ,…,OC2q2 OC31 ,OC31 ,…,OC3q3 {C4;μC4E1,μC4E2,…,μC4Ep} {C5;μC5E1,μC5E2,…,μC5Ep} {Cm;μCmE1,μCmE2,…,μCmEp} AC41,AC42,…,AC4q4 AC51,AC52,…,AC5q5 …… ACm1,ACm2,…,ACmqm OC41,OC41,…,OC4q4 OC51,OC51,…,OC5q5 OCm1 ,OCm1,…,OCmqm Class-layer

20 Fuzzy Ontology

21 Ontology-based Knowledge Management System

22 CREDIT Research Center
Located at National Cheng Kung University. Supported by Walsin Lihwa Group. ( ) Contain three main research groups. More than 10 professors and 50 Ph.D or master students.

23 CREDIT KM System (cont.)
Process Management Workflow → BPM + Web service CMMI (中小企業) Mobile Workflow Document Management Knowledge Map Q and A FAQ Personalization Semantic Search Knowledge Update

24 CREDIT KM System Meeting Management Message Management
Meeting Scheduling Meeting Notification Meeting Follow-up Message Management BBS Notification Directory Service for Message Delivery

25 Enterprise Networking Resource
Non-structured Data Internet/Intranet News/Documents XML-based E-documents CMMI-based CREDIT K.M. System Personalized Service Ontology Repository Automatic Classification Document Abstraction Workflow Intelligent Mobile Delivery On-line Tracking Personal End User Construction Meeting Scheduling Semantic Search CMMI Assistant

26 Semantic Search Service(cont.)
Human-readable HTML Machine-readable XML Machine-understandable Semantic Web with Ontology (RDF,DAML+OIL)

27 Semantic Search Service
Keyword-based search Single-word query Context query Boolean query Conceptual search Conceptual query Natural language query Semantic search Ontology-reasoning query

28 Why Semantic Search? Mass information make user confused, current search engines are not good enough. (e.g. 腦科 v.s. 電腦科學) Quality is more important than Quantity Search by "what they means" not just "what they say" The user who has no idea about domain terminologies can’t find information easily.

29 Semantic Search Service Architecture
Query processing Document Preprocessing End User WWW Repository Query CKIP Repository Information Retrieval Agent Natural Language Processing Ontology Repository Query Inference Personal Thesaurus Repository Query Personalization Index Repository Indexing and Gathering statistics XML file Repository Clustering Parsing and Transforming formats Query Results

30 Personalized Service Make a specific information service that can adapt to the behavior of each user. Provide a mechanism that can observe and analyze the browsing behavior of each user. Produce a structure with personal custom and preferences for other services using.

31 Personal Ontology … User Behavior & Browsed Content Analysis s1 s2 s3
<<History>> User Used Behavior Functionalization End User WWW Personal Log Files Repository Browsed Contents Log File Recording Engine Web Content Function of Browsing Frequency Time Fuzzy Inference <<Present>> Preference Degree Content Conceptualization & Weighting Present Browsing Concepts and Weights of Browsed Sequence of User Browsing Behavior s1 s2 s3 sn … User Behavior & Browsed Content Analysis Domain Ontology User Behavior & Browsing Content Analysis Additional Weighting of Related Sequence Concepts

32 User Behavior Analysis
In order to find out user’s favor tendency, the first job is analyzing the habitual behavior of reading. Consider two features: reading time and reading frequency. Consider reading time is related with content length, change the feature to

33 Personal Ontology Feature Data Functionalization Browsing Frequency
Time Content Length Personal Log Files Repository Browsed Contents Feature Data Data Sorting … Data Clustering Sorted Functionalization Function of Feature Feature Data Functionalization Feature Data Processing Browsing Time Function of

34 Question & Answer System
Question analysis 5W1H what, who, when, where, why, and how. Indirectly question & other YesNo question…etc. Answer analysis Question type Domain Domain knowledge

35 Question & Answer System
Ontology Where What How Answer KM PM workflow Q&A Search engine Domain Knowledge extraction & learning process Question Answering Subsystem Knowledge Extraction Subsystem Receive User query Documents Return Answer Ontology supervision Question & Answer Knowledge Base Answering process User query process User

36 Question & Answer Knowledge Base (cont.)
Domain ontology Object-oriented ontology Question ontology The knowledge of question domain To Classify and extract question Answer ontology The knowledge map of Q&A knowledge base

37 Question & Answer Knowledge Base
Alternation Rule Morphological Lexical Semantic Ontology supervision Ontology management Ontology inference

38 Document Abstraction Service
PDA Cell Phone G U I Internet OFEE Agent Document Processing Agent Retrieval Agent e-News POS Tagger (CKIP) Real-time e-News Repository Fuzzy Inference Agent Chinese Term Filter … Notebook Event Ontology Filter Chinese e-News Summary Repository Chinese e-News Ontology Summarization Agent Extracted-Event Ontology e-News Repository Chinese e-News Summary Sentence Rule Base Sentence Generation Agent

39 Meeting Scheduling Service
Meeting Host Group Calendar Data Base (GCDB) user names Meeting Scheduling Decision Support System (MSDSS) Meeting Negotiation Agent (MNA) proper time with work priority Personalized Knowledge Base (PKB) Fuzzy Inference Agent (FIA) Invitees’ Devices Desk Computer Evaluation Module Cell Phone Meeting Information Knowledge Base (MIKB) Genetic Learning Agent (GLA) Notebook IFA PDA

40 The Architecture of Fuzzy Inference Agent

41 The Flow Chart of Genetic Learning Agent
Initiation Selection Crossover Mutation Current Population New Evaluation replace Personalized Knowledge Base (PKB) Meeting Information Base (MIKB) Fuzzy Inference Agent (FIA) Start elitism

42 Workflow Service 工作 流程 部門 角色 文件 管理者 使用者 新增、刪除 修改架構 修改流程 組織 架構 人員 (人事部)
載入、更新 部門角色 簽核 發文 加簽 定義會簽 文件控管 與分析 定義職務 代理人 修改角色權限 定義角色 權限 組織設計師 流程設計師 流程管理員 工作管理員 Workflow Engine 解析流程 記錄變動 執行多條 logs DB

43 Ontology Construction

44 Automatic Construction of OO Ontology
Use object-oriented data model to represent ontologies. Follow object-oriented analysis procedure to build ontologies. Apply natural language processing technology to extract key terms from documents.

45 Automatic Construction of OO Ontology
Apply SOM clustering technology to find concepts and instances. Apply data mining technology and morphological analysis to extract attributes, operations, and associations of instances. Aggregate attributes, operations, and associations of instances to class.

46 Structure of Object-Oriented Ontology
Domain Category 1 Category 2 Category k Concept 3 Attributes 3 Operations 3 Concept 1 Attributes 1 Operations 1 Concept 2 Attributes 2 Operations 2 Concept n Attributes n Operations n Concept 4 Attributes 4 Operations 4 Class-layer Instance 3 Instance 1 Instance 2 Instance m Attributes m Operations m Instance 4 Instance-layer Association Generalization Aggregation Instance-of

47 Concepts Class and Instance

48 Domain Ontology Construction (I)
Part-of-speech Tagger (CKIP) Refining Tagging Stop Word Filter Specific Class News Documents (Training Data) Nouns Set Verbs Set Academia Sinica Balanced Corpus Segmentation Standard Dictionary Chinese Electronic Dictionary Concepts Clustering Processing Data Mining Term Analyzer Association Rule Result Concepts Set Ontology Construction Procedure Chinese Electronic Dictionary Concepts Construction Agent Operations Construction Agent Academia Sinica Balanced Corpus Attributes Construction Agent Relations Construction Agent Domain Ontology

49 Domain Ontology Construction(II)
Special Domain Documents Document Pre-processing Nouns Chinese Dictionary Concept Clustering Sentences Episode Extraction Concepts Attributes, Operations, Associations Extraction Episodes DAML+OIL Format Domain Ontology Data Flow Control Flow

50 Domain Ontology Construction(III)
Ontology Construction Agent Data Flow Nouns Repository Chinese Documents Part-Of-Speech Tagger Feature Term Pre-processor Control Flow Concept Extractor Verbs Repository Domain Term Combination Processor Episode Extractor Concepts Repository New Chinese Term Repository Episode Net Extractor Episodes Repository Domain Expert Attributes-Operation- Association Extractor Episode Net Repository HowNet Knowledge Base … Chinese Domain Ontology

51 Episode Extractor (cont.)
An episode is a partially ordered collection of events occurring together.

52 Episode Extractor The following shows an example of extraction of episode from a sentence. 德國門將卡恩贏得本屆世足賽代表最佳球員的金球獎。 POS Tagger 德國(Nc)　門將(Na)　卡恩(Nb)　贏得(VJ)　本(Nes)　屆(Nf)　世足賽(Nb)　代表(Na)　最佳(A)　球員(Na)　的(DE)　金球獎(Nb)。(PERIODCATEGORY) Stop Word Filter (德國, Nc, 1)　(門將, Na, 2)　(卡恩, Nb, 3)　(贏得, VJ, 4)　(世足賽, Nb, 5)　(代表, Na, 6)　(球員, Na, 7)　(金球獎, Nb, 8) Episode Extractor 德國(Nc)_門將(Na)_卡恩(Nb) Germany_keeper_Oliver Kahn 卡恩(Nb)_贏得(VJ)_金球獎(Nb) Oliver Kahn_took_Golden Ball

53 CMMI Ontology

54 The definition of CMMI The CMMI, Capability Maturity Model –Integrated, is a model for improving organization’s processes and ability to manage the development, acquisition, and maintenance of products of services.

55 Maturity Level 2 Process Area 1（Requirement Management）
Process Area 2（Project Planning） Process Area 3（Project Monitoring and Control） Maturity Level 2 Process Area 4（Supplier Agreement Management） Process Area 5（Measurement and Analysis） Process Area 6（Process and Product Quality Assurance） Process Area 7（Configuration Management）

56

57 CMMI Level 2 Ontology 進度監控記錄表 工作產品與工作項目清單 階段：String 工作項目名稱：String
完成百分比：Integer 未完成理由：String 工作產品編號：String 工作產品名稱：String 工作項目編號：String 工作項目名稱：String 數量：String 工作產品製作：String 估計項目/估計值：String 負責人：String 工作 項目 名稱 工作項目名稱 問題矯正措施記錄表 工作項目名稱：String 矯正之問題：String 採取之矯正措施：String 預計完成日期：String 實際完成日期：String 完成百分比：Integer 度量分析報告 完成 百分比 各表單參數之輸入：String 度量分析後結果：String

58 Specific Practice  Form
進行進度 審查 SP1.6 表單建立日期 表單編號 階段 工作項目 預計完成日期 實際完成日期 完成百分比 未完成理由 版本 填表人 說明： 1.系統自動產生項目：表單建立日期、表單編號、填表人、版本 2.階段：專案階段，例如系統分析、系統設計 3.工作項目：階段裡的工作項目，例如：系統裡的那一部分需要被監控

59 Semantic CMMI Ontology (cont.)
Domain Association Aggregation Category 1 Category 2 Category n …………… C : Concept A : Attribute C1 AC11:{VC111,VC112 ,…,VC11R1} AC12 :{VC121,VC122 ,…,VC12R2} … AC1K1:{VC1K11,VC1K12 ,…,VC1K1RK1} C2 C3 AC31:{VC311,VC312 ,…,VC31R1} AC32 :{VC321,VC322 ,…,VC32R2} … AC3K1:{VC3K11,VC3K12 ,…,VC3K1RK1} AC21:{VC211,VC212 ,…,VC21R1} AC22 :{VC221,VC222 ,…,VC22R2} … AC2K1:{VC2K11,VC2K12 ,…,VC2K1RK1} C4 AC41:{VC411,VC412 ,…,VC41R1} AC42 :{VC421,VC422 ,…,VC42R2} … AC4K1:{VC4K11,VC4K12 ,…,VC4K1RK1} C5 AC51:{VC511,VC512 ,…,VC51R1} AC52 :{VC521,VC522 ,…,VC52R2} … AC5K1:{VC5K11,VC5K12 ,…,VC5K1RK1} Cm ACm1:{VCm11,VCm12 ,…,VCm1R1} ACm2 :{VCm21,VCm22 ,…,VCm2R2} … ACmK1:{VCmK11,VCmK12 ,…,VCmK1RK1} … Class-layer

60

61

62 Document Management for CMMI
Administrator End User 支援CMMI Level 2認證之文件管理平台 差異分析工具 專案管理中心 1.P1 1.P2 文件管理中心 1.P4 訊息管理中心 企業組織 的表單 流程管理中心 1.P5 1.P3 全文檢索 自動會議排程 問答知識庫 度量分析工具 1.S1 1.S2 1.S3 1.S4 Edit Ontology Document Repository Repository

63 Measurement & Analysis
SOAP Find Measurement and Analysis Service Composed of ODBC Web Client UDDI Server PFA-WS PSPA-WS PCAA-WS PFA-WS PCAA-WS WPQA-WS PBMA-WS SSS-WS WPQA-WS Composite Services PBMA-WS PPMA-WS SSS-WS WPQAS PFA-WS WSDL PSPA-WS PCAA-WS WSDL WSDL WSDL Basic Web Services Project Planning Ontology Repository Project Monitor Ontology Repository Supplier Agreement Management Ontology Repository Process and Product Quality Assurance Ontology Repository Measurement & Analysis data

64 Project Monitor & Control
JDBC SOAP Composed of Project Monitor Service Service Web Registry Find Project Monitor Supplier Agreement Client and Control Service Management Monitor Service UDDI Process and Product Quality Monitor Service Composite UDDI Services PPMS SPMS PSIMS PCAMS PRMS PRMS PDQMS SPMS PCQMS PPMS PPMS PCQMS WSDL PCAMS PPMS WSDL PRMS PSIMS WSDL PDQMS PCQMS SPMS Basic Web Services Monitor Monitor Supplier Process and Product Project Monitor Project Agreement Management Quality Assurance and Control Repository Repository Repository Data

65 CMMI Assistant Tools

66 CMMI Assistant Tools with SIM
The purpose of CMMI Assistant Tools is to provide computation support for improving processes of software organization and managing the development, acquisition, and maintenance of products or services.

67 Architecture of CASIM F D Project Management services
Supporting services Project Management services Engineering services Process Management services 1.3.0 Service Delivery Global Working Space Sub Project 1, 2 and 4 Pre Assessment Support services Gap Analysis Supporting Organizational Process Editor services D F CM services PPQA services MA services Project Closure services PP services PMC services RD services Testing services MDA supporting services OCL supporting services UML supporting services REQM services

68 UDDI/SWDL/SOAP (Web Infrastructure)
Project Lifecycle Project Initiation Planning Project Execution Closure Requirement Specifications Project Management Plan (Estimations, Quality plan, Risk Management plan) Status Report for the Week Ending Milestone Analysis Report Noncompliance Report Project Closure Report SIM UDDI/SWDL/SOAP (Web Infrastructure) Requirement Editor & Management (RD & REQM) Service Project Planning Service Scheduling tool Meeting tool Project Quality Goal (PPQA) Generator Project Tracking Tool Monitor and Control Status Report Task & Issue Tracking Tool Project Closure Analysis Service Estimation Criteria Effort Data Template Software Metrics Historical Data Template Status Report Template Milestone Analysis Template Audit Checklist Template Closure Analysis Report Template Requirement Specifications Template Risk Management Plan Template Project Management Plan Template

69 Architecture of RD Service
Requirements Development Database Project Member Project Manager RD User Interface Product Requirements Service (SP2.1, SP2.2, SP2.3) Requirements Management Service Testing Service Configuration Management Service Customer Requirements Service (SP1.1, SP1.2) Analyze / Validate Requirements Service (SP3.1, SP3.2, SP3.3, SP3.4, SP3.5)

70 Architecture of REQM Service
Requirements Management Database Project Member Project Manager REQM User Interface Requirements Commitment Record Service (SP1.1, 1.2, 1.5) Requirements Change Management Service (SP1.3) Requirements Tracking Service (SP1.4) Requirements Development Service Project Planning Service Project Monitoring and Control Service Configuration Management Service Requirements Traceability Ontology Domain Expert

71 CMMI Ontology Extraction

72 CMMI Ontology Extraction
Domain Expert Domain Ontology … CMMI Corpus Concept Set Classified Meaningful Term Set Chinese CMMI Dictionary Meaningful Terms Document Preprocessing Mechanism Term Classifier Fuzzy Inference Mechanism CMMI Knowledge Repository CMMI Ontology … Sentence Filter Sentence Generator Sentence Path Extractor

73 Future Works Construct CMMI Ontology.
Construct CMMI-based Knowledge Management System. Apply CMMI-based Knowledge Management System to Collaborative Research Projects.

74 Q & A