Strategize a Smooth Tenant-to-tenant Migration and Copilot Takeoff
Course material: semantic decision tables for open information systems
1. Open Information Systems ( WE-DINF-13880 ) 2011 - 2012
Semantic Decision Tables
dr. Yan Tang
Yan.tang@vub.ac.be
May 11, 2011, 4:00~6:00 PM
11/05/2011 | pag. 1
3. What is a decision table?
• CSA, (1970): Z243.1-1970 for Decision Tables, Canadian Standards
Association Condition
entry
Condition
stub
Condition 1 2 3 4 5 6
Age <18 >=18,<40 >=40 <18 >=18,<40 >=40
Speak required language (s) Yes Yes Yes No No No
Action
Hire *
Train *
Action
Reject * *
entry * *
Action stub
Decision
rule
11/05/2011 | pag. 3
4. Other Decision Tools
Decision
tree
Balanced
scorecard
Bayesian
11/05/2011 | pag. 4
network
6. Decision tables in IS and
business
• Easily learned, undstandable and readable
• Concise and precise
• Clear relations of decisional alternatives
• Decision rule set
– Completeness
– Correctness
– Exclusivity
11/05/2011 | pag. 6
10. Existing Decision Table
Applications and Tools
• PROLOGA: DT editor and analyser
11/05/2011 | pag. 10
11. Importance of Group
Decision Making
“Most discussions of decision making
assume that only senior executives make
decisions or that only senior executives’
decisions matter. This is a dangerous
mistake.”
Need to involve a
– Peter Drucker community of
experts!
11/05/2011 | pag. 11
12. Semantic Decision Table
(SDT)
• SDT = Semantics + DT
– Agreements among stakeholders
– Use ontologies to store semantics
• Analysis result of (a) DT(s) designed by an acting
group (McGrath’s study)
– Community based agreements
– Role of natural language in groups Semantics in SDT is
expressed through
annotations,
commitments,
definitions, instantiation
11/05/2011 | pag. 12
13. Compare to DT, SDT …
• Can support group decision
• Can deal with large tables
• No conceptual ambiguity
• Hidden/implicit decision rules are specified
• Meta decision rules are specified
11/05/2011 | pag. 13
14. Modelling Issues
• DOGMA approach to ontology engineering (OE):
– Double articulation: ontology = ontology base (lexon base)+
commitment (Prof. R. Meersman, 1999)
– Linguistic fact oriented and scalable
– Facilitates and supports deployment
– Community grounded (agreement based)
11/05/2011 | pag. 14
15. DOGMA
• Lexon: plausible binary fact, e.g., <iPhone, rings with, is
rang with, RingTone>
• commitment
– Provides multiple views on the stored SDT lexons.
– Describes particular application views of reality, e.g.,
Each iPhone rings with AT LEAST ONE Ring Tone.
– Needs to be expressed by a commitment language
e.g. p1 = [iPhone, rings with, is rang with, RingTone]:
MAND (p1).
11/05/2011 | pag. 15
16. Use ORM to Graphically
Model Commitments
• ORM – a method for designing models at
the conceptual level, where the application
is described in terms easily understood by
non-technical users.
11/05/2011 | pag. 16
17. An SDT Example
Condition 1 2 3 4
People move Ear Yes No Yes No
Pressure on Crib Yes Yes No No
Action
Screen shows Message Message1
iPhone rings RingTone1
SDT Lexons
Lexon 1 <Bunny, has, is of, Ear>
Lexon 2 <Bunny, has, is of, Name>
Lexon 3 <Ear, is moved by, move, People>
Lexon 4 <Crib, has, is of, Name>
Lexon 5 <Screen, shows, is shown by, Message>
Lexon 6 <iPhone, rings with, is rang with, RingTone>
SDT Commitments
Commitment 1 EACH Bunny has EXACT ONE name.
Commitment 2 EACH Crib has EXACT ONE name.
Commitment 3 EACH Screen shows AT LEAST ONE Message
Commitment 4 Each iPhone rings with AT LEAST ONE Ring Tone.
Instantiation of Decision Items
People move Ear “People” is James. “Ear” is the ear from the Bunny in the living room.
Pressure on Crib “Crib” is James’ crib. “Pressure on Crib – Yes” means that James is in his crib.
Screen shows Messages “Screen” is the smart screen in the living room.
iPhone rings “iPhone” is Mary’s iPhone. She has only one iPhone.
11/05/2011 | pag. 17
18. How to Construct an SDT?
Study decision Define Scope
maker environment decision To study the
To form decision individuals input problems
decision maker
group and define
candidates and
decision tasks
problem
background
Define
Define a Define
thresholds for
group decision tasks
group behaviour
To create
decision tables
and make into
Design Extract SDT
Design SDT
decision SDT
commitments
table lexons
11/05/2011 | pag. 18
19. What is Inside an SDT
Commitment?
Constraint types:
Uniqueness
Mandatory
Occurrence frequency Operators:
Negation
Subset
Conjunction
Equality Disjunction
Exclusion Implication
Value constraint
Sequence
Subtype
Other dependencies
11/05/2011 | pag. 19
20. Be Careful!
• Top common supertype.
– It is impossible to have an object type that has two (or more than two)
mutually exclusive super types, unless the instance set of this type is
empty.
• Exclusion-mandatory.
– If an object type plays a mandatory role and it participates in an
exclusion constraint with other roles, then only the role that connects to
both the mandatory constraint and the exclusion constraint is played.
Other roles are never played.
• Set-comparison constraints.
– It is impossible to apply both the subset constraint and the exclusion
constraint at the same time to two lexons.
• …
11/05/2011 | pag. 20
21. Use Ontology to Make a Good Semantic Decision Table
SEMANTIC DECISION TABLE –
ANALYSIS (PART I)
11/05/2011 | pag. 21
22. Validation and Verification
• In order to make a “good” SDT, it needs to
be validated and verified
11/05/2011 | pag. 22
23. SDT Analyser (Part I)
• Completeness Validation and
verification
• Entries Validity issues
• Transaction Test
• Identification of Impossible Rules
• Identification of Overlapping Rules
• Identification of Rule Gaps
11/05/2011 | pag. 23
24. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser –
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Completeness
Completeness
11/05/2011 | pag. 24
25. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser –
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Completeness
11/05/2011 | pag. 25
26. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Entries
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Validity
• Example 1 – invalid Boolean stub
11/05/2011 | pag. 26
27. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Entries
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Validity
• Example 2 – Invalid Set Entry
11/05/2011 | pag. 27
28. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Entries
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Validity
• Example 3 – Invalid Float Entries
11/05/2011 | pag. 28
29. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Entries
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Validity
• Example 4 – Invalid Integer Entries
11/05/2011 | pag. 29
30. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Transaction
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Test
11/05/2011 | pag. 30
31. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Horizontal
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Impossible Rules
• Example 1 – Condition Entry is a String
Set
What if
X=Y?
11/05/2011 | pag. 31
32. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Horizontal
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Impossible Rules
• Example 2 – Using Implication Operator in
the SDT Commitment
P1 = [CONDITION, has, is of,
CONDIT ION_ENTRY]: IMPLIES (P1
( C O N D I T I O N _ E N T RY ) = ” ” , O R ( ( P 1
(CONDITION_ENTRY) =”J”), P1
(CONDITION_ENTRY) =”N”)).
11/05/2011 | pag. 32
33. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Vertical
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Impossible Rules
• Example 1 – use constraints of
equivalence and value range
11/05/2011 | pag. 33
34. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Vertical
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Impossible Rules
• Example 2 – use constraints of
equivalence and set
11/05/2011 | pag. 34
35. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser – Vertical
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Impossible Rules
• Example 3 – use constraints of exclusion
11/05/2011 | pag. 35
36. • Completeness
SDT Analyser –
• Entries Validity
•
•
•
Transaction Test
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Overlapping
• Identification of Rule Gaps
Rules
• Example 1 – Value
What If
C1=C2?
11/05/2011 | pag. 36
37. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser –
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Identification of Rule Gaps
• Example 1
11/05/2011 | pag. 37
38. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser –
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Identification of Rule Gaps
• Example 2
11/05/2011 | pag. 38
39. •
•
•
Completeness
Entries Validity
Transaction Test
SDT Analyser –
•
•
•
Identification of Impossible Rules
Identification of Overlapping Rules
Identification of Rule Gaps
Identification of Rule Gaps
• Example 3
11/05/2011 | pag. 39
40. Use Ontology to Make a Good Semantic Decision Table
SEMANTIC DECISION TABLE –
ANALYSIS (PART II)
11/05/2011 | pag. 40
41. SDT Analyser –
Unreachable Rules
• Example 1 – condition entries are sets
11/05/2011 | pag. 41
42. Applied constraint types –
Subtyping
• P = [Lecturer, is a, is, Teacher]: SUBTYPE (P).
11/05/2011 | pag. 42
47. Applied Constraint Types –
Exclusive-Or
• P = [TABLE, has, is of, ACTION]: OR (P (ACTION) = “Accept”, P
(ACTION) = “Refuse”).
• (P6 = [Lecturer, is a, is, Person], P7 = [Driver, is a, is, Person]): OR
(P6 (is), P7 (is)).
11/05/2011 | pag. 47
48. Other Extensions to
Decision Tables
• Second-Order Decision Tables (SODT)
• Fuzzy Decision Tables (FDT)
• Rough Set Decision Tables (RSDT)
11/05/2011 | pag. 48
49. Conclusion
• What is SDT?
• How to build an SDT?
• How to build a good SDT?
11/05/2011 | pag. 49