Unit 3

You might also like

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 22

PAT trees and PATArrays

Courtesy:
Modern Information Retrieval
by R. Baeza-Yates and B. Ribeiro-Neto

1
Introduction

 Word-based indexing
» Inverted indices are good for search words
» Queries such as phrases are expensive to solve using
Inverted files
» For word-based applications, inverted files perform better
 Suffix trees and suffix arrays
» complex queries

2
Text Suffixes

This is a text. A text has many words. Words are made from letters.

text. A text has many words. Words are made from letters.
text has many words. Words are made from letters.
many words. Words are made from letters.
words. Words are made from letters.
Words are made from letters.
made from letters.
letters.

3
The Suffix Trie and Suffix Tree

1 11 19 28 33 40 46 50 60
This is a text. A text has many words. Words are made from letters.

w l
t m
60 6 5 3 60

o e a
28 50

r n d
x
28 50 11 19

d
t
33 40

s
11 19

33 40 4
PAT Trees and PAT Arrays
Information Retrieval: Data Structures and Algorithms
by W.B. Frakes and R. Baeza-Yates (Eds.)
Englewood Cliffs, NJ: Prentice Hall, 1992.
(Chapters 5)

5
PAT Trees and PAT Arrays

 Problems of tradition IR models


» Documents and words are assumed.
» Keywords must be extracted from the text (indexing).
» Queries are restricted to keywords.
 New indices for text
» A text is regarded as a long string.
» Each position corresponds to a semi-infinite string (sistring).
» No structures and no keywords

6
Semi-infinite Strings

 Example
Text Once upon a time, in a far away land …
sistring 1Once upon a time …
sistring 2nce upon a time …
sistring 8on a time, in a …
sistring 11 a time, in a far …
sistring 22 a far away land …
 Compare sistrings
22 < 11 < 2 < 8 < 1

7
PAT Tree

 PAT Tree
A Patricia tree constructed over all the possible sistrings of a text
 Patricia tree
» a binary digital tree where the individual bits of the keys are used to
decide on the branching
– A zero bit will cause a branch to the left subtree
– A one bit will cause a branch to the right subtree
» each internal node indicates which bit of the query is used for
branching
– absolute bit position
– a count of the number of bits to skip
» each external node points to a sistring
– the integer displacement to original text

8
Example 1

Text 01100100010111 … 2 2
sistring 1 01100100010111 …
4 1 3 2
sistring 2 1100100010111 …
sistring 3 100100010111 … 1
sistring 4 00100010111 …
2 2
sistring 5 0100010111 …
sistring 6 100010111 … 4 3 3 2
sistring 7 00010111 …
sistring 8 0010111 ... 5 1

: external node sistring


1 1 1 (integer displacement)
0 1 0 1
1 1 total displacement of the
2 2
0 1 bit to be inspected
3 2 : internal node
skip counter & pointer
1
Text 01100100010111 …
sistring 1 01100100010111 … 2 2
sistring 2 1100100010111 …
sistring 3 100100010111 … 3 3 4 2
sistring 4 00100010111 …
sistring 5 0100010111 … 7 4 5 1 6 3
sistring 6 100010111 …
sistring 7 00010111 … 1
sistring 8 0010111 ...
1 2 2
3 3 4 2
2 2
4 7 5 5 1 6 3
4 3 2
4 8
5 1 6 3
Search 00101
註: 3 和 6 要 4 個 bits 才能區辨
Indexing Points

 The above example assumes every position in the text is


indexed.
i.e. n external nodes, one for each indexed position in the text
 Word and phrase searches
sistrings that are at the beginning of words are necessary
 Trade-off between size of the index and search requirements

11
Prefix searching

 idea
every subtree of the PAT tree has all the sistrings with a given
prefix.
 Search: proportional to the query length
exhaust the prefix or up to external node.

Search for the prefix


“10100” and its answer

12
Proximity Searching

 Find all places where s1 is at most a fixed (given by a user)


number of characters away from s2.
in 4 ation ==> insulation, international, information
 Algorithm
1. Search for s1 and s2.
2. Select the smaller answer set from these two sets and
sort by position.
3. Traverse the unsorted answer set, searching every
position in the sorted set and checking if the distance
between positions satisfying the proximity condition.

sort+traverse time:m1 logm1 +m2logm1 (assume m1<m2)

13
Range Searching

 Search for all the strings within a certain lexicographical range.


» Ex: the range of “abc” ..”acc”:
– “abracadabra”, “acacia” ○
– “abacus”, “acrimonious” X
 Algorithm
» Search each end of the defining intervals.
» Collect all the sub-trees between (and including) them.

14
Longest Repetition Searching

 the match between two different positions of a text where this match is
the longest in the entire text, e.g.,
01100100010111 the tallest internal node gives a pair
of sistrings that match for the greatest
number of characters
Text 01100100010111
sistring 1 01100100010111 1
sistring 2 1100100010111
2 2
sistring 3 100100010111
sistring 4 00100010111 3 3 4 2
sistring 5 0100010111
sistring 6 100010111 7 5 5 1 6 3
sistring 7 00010111
sistring 8 0010111 4 8
15
“Most Significant” or “Most Frequent” Matching

 The most frequently occurring strings within the text database


» e.g., the most frequent trigram
 Find the most frequent trigram
» find the largest subtree at a distance 3 characters from root

1 the tallest internal node gives a pair


of sistrings that match for the greatest
2 2 number of characters

3 3 4 2 i.e., 1, 2, 3 are the same for


sistrings 100100010111
7 5 5 1 6 3 and 100010111

4 8
16
Building PAT Trees as Patricia Trees (1)

 Bucketing of external nodes


» collect more than one external node
» a bucket replaces any subtree with size less than a certain
constraint (b)
save significant number of internal nodes
» the external nodes inside a bucket do not have any structure
associated with them
increase the number of comparisons for each search

17
Building PAT Trees as Patricia Trees (2)

 Mapping the tree onto the disk using super-nodes


» Advantage: save the number of disk access and space
» Every disk page has a single entry point, contains as much of the
trees as possible, and
– terminates either in external nodes or in pointers to other disk pages
– The pointers in internal nodes will address either a disk page or another
node inside the same page
 reduces the storage cost of internal nodes
» Example
– Assume a disk page contains on the order of 1,000 internal/external
nodes
– on the average, each disk page contains about 10 steps of a root-to-
leaf path

18
PAT Trees Represented as Arrays

 External node bucket size, b


 If we keep the external nodes in the bucket in the same relative order
as they would be in the tree
» Indirect binary search vs. sequential search
PAT array
1
7 4 8 5 1 6 3 2
2 2

3 3 4 2
0 1 1 0 0 1 0 0 0 1 0 1 1 1 ...
7 Text
5 5 1 6 3

4 8
19
Searching PAT Trees as Arrays

 Prefix searching and range searching


doing an indirect binary search over the array with the results of the
comparisons being less than, equal, and greater than.
 Example
Search for the prefix 100 and its answer
 Most frequent, Longest repetition
» Manber and Baeza-Yates (1991)

PAT array
7 4 8 5 1 6 3 2

0 1 1 0 0 1 0 0 0 1 0 1 1 1 ...
Text 20
Comparisons

 Signature files
» Use hashing techniques to produce an index
» Advantage
– storage overhead is small (10%-20%)
» Disadvantages
– the search time on the index is linear
– some answers may not match the query, thus filtering must be
done

21
Comparisons (Continued)

 Inverted files
» storage overhead (30% ~ 100%)
» search time for word searches is logarithmic
 PAT arrays
» potential use in other kind of searches
– phrases
– regular expression searching
– approximate string searching
– longest repetitions
– most frequent searching

22

You might also like