Graph Model for Pattern Recognition in Text - Mining and Analyzing Social Networks - page 4

2.2.1 Record the Keyword Information Appeared in the
Document

For a given document, the following steps are applied to it. Suppose we
have already chosen a set of words as keywords, say

K { K 1 , K 2 , ··· , K m }

.
Record every keyword and its position in the document. We will use the
following notation:
•

=

k i represents one of the keyword in the keyword set K .

• i represents the order of the keyword appearing in the document. (It is

possible that k i ,k j

are the same element in K .)

• m represents the total number of keywords appearing in the document.

• p i is a integer, which represents the total number of the words from the

beginning of the document to the word k i .
In addition we record the frequency of each keyword at the same time. Thus
we have the Keyword-Position information table(Table 1).

Table 1 Keyword-Position table

Keyword appears
in the document

Position
in the document

k 1 p 1

k 2 p 2

. .

k m p m

The details of this process are illustrated as follows (with Figure 1 as an
example). For this example, we use the keyword set: { bank, fund, account,
transfer } . Its Keyword-Position information is listed in Table 2. Frequency
information of each keyword for the given example (Figure 1) is listed in
Table 3.

Table 2 Keyword-Position information of the email inFigure 1

Keyword appears
in the document

Position
in the document

bank 91

fund 103

account 109

transfer 124

fund 153

transfer 155

account 158