Question

Positional Indexes: Example Query: "to, be2 or3 not4 tos be То, 993427: (1: (7, 18, 33, 72, 86, 231): 2: (1, 17, 74, 222, 255); 4: (8, 16, 190, 429, 433): 5: (363, 367); 7: (13, 23, 191); ...) BE, 178239: 4: (17, 191, 291, 430, 434); 5: (14, 19, 101); ...)

what is the positional index and I want you to solve this example please in information Retrieval ?

Answer 1

Positional indexes are a more efficient way for information retrieval. we store postings of the form docID: <position1, position2, Position3....

the query is 'to be or not to be'. The postings lists is: to, be, or, not. We have to intersect the postings lists for to and be. We need to first look for documents that contains both. After that we look for places in the document where there is occurrence of be with index one higher than position of to .In the above list given in picture, the pattern of occurrences that is a possible match is 4: <8, 16, 190, 429, 433> and 4: <17, 191, 291, 430, 434>

index of 'be' is one higher than index of 'to'

index of 'to'....16

index of 'be'...17

index of 'to'....190

index of 'be'...191

index of 'to'....429

index of 'be'...430

index of 'to'....433

index of 'be'...434

In the given example the meaning is the word "TO" has a document frequency 993427 and in document 1 it occurs at positions

7,18,33,72,86,231 i.e occurance of to is 6 times in document 1.

TO, 993427:

1:<7, 18, 33, 72, 86, 231>

2: <1, 17, 74, 222, 255>

4: <8, 16, 190, 429, 433>

5: <363, 367>

7: <13, 23, 191>

BE, 178239:

   1: <17, 25>

4: <17, 191, 291, 430, 434>

5: <14, 19, 101>

Query: “to1 be2 or3 not4 to5 be6”

we want to search "to be" so we need to progressively intersect the two documents given

in both the documents we have document 1

1:<7, 18, 33, 72, 86, 231> and

1: <17, 25>

when we merge these we don't have a match so we discard that document.

Document number 2 is not there in the 'Be' list .so we discard that also. doc id 2 is not present in the 'be' list

then we take document number 4 from both the list

from 'to' list we have 4: <8, 16, 190, 429, 433>

and from 'Be ' list we have 4: <17, 191, 291, 430, 434>

when we look at this doc id matches.Both have doc id=4. Then we have to merge them.when look at both we could find a match

$kuWXsvgAAAABJRU5ErkJggg==$

Matches in both documents are underlined with colors.we have three matches in the document 4. Hence we may conclude that Document 4 is a match.

4: <8, 16, 190, 429, 433 4: <17,191, 291, 430, 434>