Annotation Free Information Extraction

1
Annotation Free Information Extraction

• Chia-Hui Chang
• Department of Computer Science Information
  Engineering
• National Central University
• chia_at_csie.ncu.edu.tw
• 10/4/2002

2
IEPAD Information Extraction based on Pattern
Discovery

• C.H. Chang.
• National Central University
• WWW10

3
Semi-structured Information Extraction

• Information Extraction (IE)
• Input Html pages
• Output A set of records

4
Pattern Discovery based IE

• Motivation
• Display of multiple records often forms a
  repeated pattern
• The occurrences of the pattern are spaced
  regularly and adjacently
• Now the problem becomes ...
• Find regular and adjacent repeats in a string

5
IEPAD Architecture
6
The Pattern Generator

• Translator
• PAT tree construction
• Pattern validator
• Rule Composer

7
1. Web Page Translation

• Encoding of HTML source
• Rule 1 Each tag is encoded as a token
• Rule 2 Any text between two tags are translated
  to a special token called TEXT (denoted by a
  underscore)
• HTML Example
• ltBgtCongolt/BgtltIgt242lt/IgtltBRgt
• ltBgtEgyptlt/BgtltIgt20lt/IgtltBRgt
• Encoded token string
• T(ltBgt)T(_)T(lt/Bgt)T(ltIgt)T(_)T(lt/Igt)T(ltBRgt)
• T(ltBgt)T(_)T(lt/Bgt)T(ltIgt)T(_)T(lt/Igt)T(ltBRgt)

8
Various Encoding Schemes
9
2. PAT Tree Construction

• PAT tree binary suffix tree
• A Patricia tree constructed over all possible
  suffix strings of a text
• Example
• T(ltBgt) 000
• T(lt/Bgt) 001
• T(ltIgt) 010
• T(lt/Igt) 011
• T(ltBRgt) 100
• T(_) 110

T(ltBgt)T(_)T(lt/Bgt)T(ltIgt)T(_)T(lt/Igt)T(ltBRgt) T(ltBgt)T(
_)T(lt/Bgt)T(ltIgt)T(_)T(lt/Igt)T(ltBRgt)
000110001010110011100 000110001010110011100
10
The Constructed PAT Tree
11
Definition of Maximal Repeats

• Let a occurs in S in position p1, p2, p3, , pk
• a is left maximal if there exists at least one
  (i, j) pair such that Spi-1?Spj-1
• a is right maximal if there exists at least one
  (i, j) pair such that Spia?Spja
• a is a maximal repeat if it it both left maximal
  and right maximal

12
Finding Maximal Repeats

• Definition
• Lets call character Spi-1 the left character
  of suffix pi
• A node ? is left diverse if at least two leaves
  in the ?s subtree have different left characters
• Lemma
• The path labels of an internal node ? in a PAT
  tree is a maximal repeat if and only if ? is left
  diverse

13
3. Pattern Validator

• Suppose a maximal repeat ? are ordered by its
  position such that suffix p1 lt p2 lt p3 lt pk,
  where pi denotes the position of each suffix in
  the encoded token sequence.
• Characteristics of a Pattern
• Regularity Variance coefficient
• Adjacency Density

14
Pattern Validator (Cont.)

• Basic Screening
• For each maximal repeat a, compute V(a) and D(a)
• a) check if the patterns variance V(a) lt 0.5
• b) check if the patterns density 0.25 lt D(a) lt
  1.5

15
4. Rule Composer

• Occurrence partition
• Flexible variance threshold control
• Multiple string alignment
• Increase density of a pattern

16
Occurrence Partition

• Problem
• Some patterns are divided into several blocks
• Ex Lycos, Excite with large regularity
• Solution
• Clustering of the occurrences of such a pattern

Clustering
V(P)lt0.1
No
P
Discard
Yes
Check density
17
Multiple String Alignment

• Problem
• Patterns with density less than 1 can extract
  only part of the information
• Solution
• Align k-1 substrings among the k occurrences
• A natural generalization of alignment for two
  strings which can be solved in O(nm) by dynamic
  programming where n and m are string lengths.

18
Multiple String Alignment (Cont.)

• Suppose adc is the discovered pattern for token
  string adcwbdadcxbadcxbdadcb
• If we have the following multiple alignment for
  strings adcwbd'', adcxb'' and adcxbd''
• a d c w b d
• a d c x b -
• a d c x b d
• The extraction pattern can be generalized as
  adcwxbd-

19
Pattern Viewer

• Java-application based GUI
• Web based GUI
• http//www.csie.ncu.edu.tw/chia/WebIEPAD/

20
The Extractor

• Matching the pattern against the encoding token
  string
• Knuth-Morris-Pratts algorithm
• Boyer-Moores algorithm
• Alternatives in a rule
• matching the longest pattern
• What are extracted?
• The whole record

21
Experiment Setup

• Fourteen sources search engines
• Performance measures
• Number of patterns
• Retrieval rate and Accuracy rate
• Parameters
• Encoding scheme
• Thresholds control

22
Translation

• Average page length is 22.7KB

23
Accuracy and Retrieval Rate
24
Problems

• Guarantee high retrieval rate instead of accuracy
  rate
• Generalized rule can extract more than the
  desired data
• Only applicable when there are several records in
  a Web page, currently

25
ROADRUNNER Towards Automatic Data Extraction
from Large Web Sites

• Valter Crescenzi , Giansalvatore , Paolo Merialdo
• VLDB2001

26
Observations

• 1. Wrapper generator works by using additional
  information. (labeled samples)
• 2. Wrapper induction system has some a priori
  knowledge about the page organization.
• 3. Finally, systems generate wrapper by examining
  one HTML page at a time.

27
ROADRUNNER new perspective

• 1. Dont rely on any interaction with the user.
  (Completely automatic)
• 2. No a priori knowledge
• HTML schema will be inferred along with wrapper.
• Can handle any nested structures.
• 3. Works with two HTML pages at a time. (based on
  the study of similarities and dissimilarities
  between the pages)

28
(No Transcript)
29
Theoretical Background

• Site generation Encoding of database content
• Data extraction Decoding
• The problem is based on a close correspondence
  between nested type and union-free regular
  expressios.

30
Delimiter

• PCDATA map to string
• map to lists (nested) , being iterator
• ? map to nullable fields, optional patterns.
• Find schema and data extraction Find minimal
  UFRE.

31
Matching Technique

• It is based on a matching technique called ACME.
  (Align, Collapse under Mismatch, and Extract)
• HTML ? XHTML ? tokens
• Matching algorithm works on two objects
• A list of tokens, call the sample
• A wrapper (one UFRE)
• This is done by solving mismatches between the
  wrapper and the sample.

32
(No Transcript)
33
Mismatches

• 1. String mismatches
• May be due only to different values of a database
  field.
• These mismatches are use to discover fields.
  (PCDATA)
• Ex John Smith and Paul Jones at token 4
• 2. Tag mismatches
• Optional patterns
• Iterative patterns

34
Discovering Optionals

• Strategy Looking for repeated patterns as a
  first step, and then, if this attempt fails, in
  trying to identify optional pattern.
• Two steps
• 1. Optional Pattern Location by Cross-Search
• Mismatch at token 6 - ltULgt and ltIMG/gt
• Assume optional pattern is located on wrapper or
  sample.
• 2. Wrapper Generalization
• ( ltIMG src/gt ) ?

35
Discovering Iterators

• 1. Square Location by Terminal Tag Search
• Both the wrapper and sample contain at least one
  occurrence of the square.
• Terminal Tag position before the mismatch
• In this example is lt/LIgt
• Test which is the square initial tag ?
• lt/UIgt lt/LIgt v.s. ltLIgt lt/LIgt
• Finally, we can infer that the sample contains
  one candidate occurrence of the square at token
  20-25.

36
Discovering Iterators (cont)

• 2. Square Matching
• Try to match the candidate square occurrence
  (tokens 20-25).
• Backwards matching token 25 and 19, then moves
  to 24 and 18 and so on.
• 3. Wrapper Generalization
• If we denote the newly found square by s, we
  replace the repeated pattern by (s)

37
More Complex Example

• First mismatch at token 15 (external mismatch)
• Find iterators
• Terminal tag lt/LIgt
• Candidate square is found ltLIgt lt/LIgt at token
  15-28
• Backward match second mismatch at token 23 and
  9 (internal mismatch) ? solve the mismatch by
  recursive

38
Recursively solve mismatch

• Internal mismatch at token 23 and 9
• Solve it by the same way at external mismatch.
• But dont work by comparing one wrapper and one
  sample, rather two different portions of the same
  objects.
• Terminal tag ltBgt
• Candidate square is lt/BgtltBgt token 23-18
• Backward match mismatch at token 20 and 26
• Find token 20-22 is optional pattern.

39
(No Transcript)
40
Matching as an AND-OR tree

• Finding one solution to match(w,s) corresponds to
  finding one visit for the AND-OR tree.
• (i) match(w,s) all external mismatches
  encountered during the parsing (AND node)
• (ii) solve mismatch by either introducing one
  field, or one iterator, or one optional (OR)
• (iii) The search may either on wrapper or sample
  (OR)
• (iv) iterators and optionals are various
  candidates (OR)
• (v) Discover iterators may be need to recursively
  solve several internal mismatches. (AND)

41
AND-OR tree
42
Experimental Results
43
Experimental Results (cont)
44
Extracting Structured Data from Web Page

• Arvind Arasu, Hector Garcia-Molina
• ACM SIGMOD 2003

45
Cue

• Keywords schema, template
• Web pages belonging to the same site are
  generated by encoding data of the same schema
  with a common template
• gt a common template by plugging-in value

46
Figuration
47
Goal and Challenge

• Previous IE Techniques rely on heuristic by
  human. ex. wrapper
• Goal to deduce the template without human
• Time consuming and error-prone
• Optional attributes are ignored
• Challenge
• No obvious way of differentiating what text is
• template or data
• The schema of data in pages isnt flat but more
• complex and semi-structured of attributes

48
Model, Problem Formulation

• Structured Data
• Model of Page Creation
• Optionals and Disjunctions
• Problem Statement
• Miscellaneous Terminology, Definition

49
Structured Data

• Token A token is some basic unit of text
• Structured Data any set of data values
  conforming to a common schema or type
• Define Type
• 1. Basic Type (ß) string of tokens
• e.g. lthtmlgt, text
• 2. Ordered List Type tuple constructor order n
• e.g. ltT1, T2, , Tngt, T1, T2, , Tn type
• 3. Define Type set constructor
• e.g. T , T type

50
Define term value and example

• Define instance
• 1. an instance of basic type, ß, token
• 2. an instance of type ltT1, T2, , Tngt is
• tuple of the form lti1, i2, , ingt, attributes
• i1, i2, , in are instances of typesT1,
  T2, , Tn
• 3. an instance of type T, is any set of
  elements
• e1, e2, , em, such ei is an instance of
  type T
• Instance ? Value String ? token
• Example
• Schema S1
• Value

51
Model of Page Creation

• Definition A template T for a schema S (as shown
  TS), is defined as a function that maps each type
  constructor t of S into an ordered set of strings
  T(t ), such that,
• tis the tuple constructor of order n, T(t) is an
  order set of n1 string
• tis the set constructor of order n, T(t) is
  string St

?(T, x) values x that are instances of
sub-schema of S
52
Encoding of a value x? S

• 1. if x ?ß, then ? (T,x)?x
• 2. if x ? ltx1, x2, , xngttt
• ? (T,x) ? C1 ? (T, x1) C2 ? (T, xn) Cn1
• 3. if x ? e1, e2, , emts , ts ? S
• ? (T,x) ? ? (T, e1) S ? (T, e2) .S ? (T, em)

53
Example of Schema S1
54
Optionals and Disjunctions

• Optional
• If T is type, optional type (T)?Tt
• t 0 or 1
• Disjunction
• If T1 and T2 is type, disjunction type
• (T1 T2) ltT1t1, T2t2 gtt
• t1t2 1

55
Problem Statement

• Extract Problem n pages, pi ?(T, xi)
• (1 i n), created from some unknown
  deduction template T and values x1,. . .,x1
  from the set of pages alone

56
Example of correct solution of EXTRACT (cont.)
57
Example of correct solution of EXTRACT (cont.)
58
Miscellaneous Terminology, Definition

• An occurrence of a token in template is called a
  template-token
• An occurrence of a token in value is called a
  value-token
• An occurrence of a token in page is called a
  page-token
• 2 page-token in Pe have the same role iff they
  have been generated by the same template-token

59
Overview Approach - EXALG
(ECGM)
60
EXALG - ECGM FINDEQ (step2)

• The module used to compute equivalence
  classese, set of tokens having the same
  frequency of occurrence in every pages Pe
• Ex. ee1 lthtmlgt, ltbodygt, Book, Reviews, ltolgt,
• lt/olgt, lt/bodygt, lt/htmlgt
• Ex. ee3 ltligt, Reviewer, Rating, Text, lt/ligt
• EXALG retain only EQ Classes that are Large and
  Frequently occurring EQ Classes (LFEQ)

61
EXALG - ECGM HANDINV (step3)

• The module used to detect and remove invalid
  LFEQs those that are not formed by tokens
  associated with a type constructor

62
DIFFFORM (step1) and DIFFEQ (step4)

• The module used to add more tokens to LFEQ by
  differentiating roles
• Ex. Name has multiple role, one occurs in Book
  Name and the other occurs in Reviewer Name
• Differentiate the multiple roles
• The multiple tokens occur in different path from
  root in the HTML parse tree (DIFFFORM)
• The multiple tokens occur in different Position
  with respect to LFEQ ee1(DIFFEQ)
• dtoken ex. Name5 and Name14
• regard NameA and NameB as different tokens

63
Review ECGM
64
Example After ECGM Process

• ee1 lthtmlgt, ltbodygt, ltbgt, Book, Name, lt/bgt,
  ltbgt, Reviews, lt/bgt, ltolgt, lt/olgt, lt/bodygt, lt/htmlgt
  
• 8 ?13
• ee3 ltligt, ltbgt, Reviewer, Name, lt/bgt, ltbgt,
  Rating, lt/bgt, ltbgt, Text, lt/bgt, lt/ligt
• 5 ?12
• Position empty and non-empty

65
Construct Schema from ECGM

• Construct Schema S fromee1
• The 1st of non-empty position is Basic Type ß
• The 2nd of non-empty position is ee3 , are
  generated by set type constructorte3
• ? T(te1) ltC11, C12,C13gt, S ltß, S te2
  gtte1
• ? T(te2) S lt C31, C32,C33,C34 gt
• ? T(te3) lt C31, C32,C33,C34 gt, ltß,ß,ß,gtte3
• ? S lt ß, ltß,ß,ß,gtte3 te2 gtte1

66
Equivalence Classes (Cont.)

• Pages P p1, , pn , pi ?(TS, xi)
• TS t1, , tk type constructor
• Definition All tokens of equivalence class have
  the same occurrence vector
• ex. ee1 lt1,1,1,1gt ee3 lt1,2,1,0gt
• Observation1 Tokens associated with the same
  type constructor tj in T that have unique-roles
  occur in the same equivalence class. (used to
  decide EQ valid or not)
• Support of token (page contain)
• Size of EQ class (token of EQ)

67
Equivalence Classes (Cont.)

• Observation2 for real pages, an equivalence
  class of large size and support is usually valid
• Properties of EQ class ltt1, , tmgt
• Ordered
• Nested the span of all occurrences of ei is
  within for some fixed Position_p or doesnt
  overlap
• Observation3 A valid equivalence class is
  ordered and a pair of two valid equivalence
  classes is nested

68
Handling Invalid Equivalence classes

• Detect the existence of invalid LFEQs using
  violation of ordered and nesting
• Yes, discard some of LFEQs and break other into
  smaller LFEQs

Differentiating roles of tokens

• By Path different roles of tokens are in
  different path of HTML parse tree
• By Position different roles of tokens locates
  at different Position (non-empty)

69
Equivalence Class Generation Module

• OUTPUT set of LFEQs of dtokens and page
  represented as string of dtokens
• FINDEQ 2 parameters used to consider
• LFEQs (SIZETHRES, SUPTHRES)
• On running example
• SIZETHRES SUPTHRES 3
• the iteration 2, find out ee1 and ee3

70
Building Template and Extracting Values

• Input to this module is e1 ,e2 , ,em
• The ANALYSIS consist of 2 modules CONSTTEMP and
  EXVAL
• CONSTTEMP ,ei d1, d2, , dl
• Start the basic e1 lthtmlgt, ltbodygt, ,lt/bodygt,
  lt/htmlgt
• recursively constructs a template Tei ,
  corresponding toei , and template Tei, p,
  corresponding to each non-empty position p ofei
• Checks if the set of strings, PosString(ei ,p),
  corresponding has some recognizable pattern

71
Example

• In running example, PosString(ee1 ,6) is a
  string dtokens for every occurrence of ee1,
  which matches Pattern 5 of table PosString(ee1
  ,10) is always a string of 0 or more occurrences
  of ee3, which matches Pattern 1
• ee1 lthtmlgt, ltbodygt, ltbgt, Book, Name, lt/bgt,
  ltbgt,
• Reviews, lt/bgt, ltolgt, lt/olgt,
  lt/bodygt, lt/htmlgt

72
Assumption

• The 4 assumptions
• (A1) A large number of tokens occurring in
• template have unique roles
• (A2) The EQ class derived from a type constructor
• is recognized as an LFEQ
• (A3) Irregularity in encoded data that leads to
• invalid EQ class
• (A4) The separators are around data values. In
• this model, strings associated with type
• construction are non-empty position

73
Evaluation

• Leaf attribute Am in schema Sm
• Correct the set of Am in the page is equal to
  the set of extracted value Ae in the page
• Partially Correct the set of Am in the page is
  not equal to the set of extracted value Ae in the
  page, but as part of value of Ae
• Incorrect not correct and Partially correct

74
Result

• 18 or 40 of input collections our System
  correctly extracted all the attribute
• Around 80 of the attributes were extracted
  correctly
• Normalized average
• Input size lt10
• Parameter 3

75
Conclusion

• EXALG use 2 novel concept equivalence classes
  and differentiate roles, to discovery the
  template
• Impact of the failed assumption is limit to a few
  attributes
• Future work
• Develop techniques for crawling, indexing, and
  providing querying support for the structured
  pages in the web
• Develop techniques for automatically annotating
  the extracted data, possibly using the words that
  appear in the template

76
References

• C.H. Chang. and S.C. Lui. IEPAD Information
  Extraction based on Pattern Discovery, WWW2001,
  pp. 681-688.
• Valter Crescenzi, Giansalvatore Mecca, Paolo
  Merialdo. RoadRunner Towards Automatic Data
  Extraction from Large Web Sites. VLDB2001,
  109-118
• Arvind Arasu, Hector Garcia-Molina. Extracting
  Structured Data from Web Pages. SIGMOD2003,
  337-348.