Abstract
In this paper, we propose TFIGF, a method which detects peculiar web pages using distribution of words in WWW given a set of keywords. Our TFIGF detects a set of index words which represent a WWW page by estimating their importance in the WWW page and their rareness in WWW. Experiments using both English and Japanese WWW pages clearly show superiority of our approach over a traditional method which employs a limited number of WWW pages in the estimation.
| Original language | English |
|---|---|
| Pages (from-to) | 355-362 |
| Number of pages | 8 |
| Journal | Lecture Notes in Computer Science |
| Volume | 3245 |
| Publication status | Published - 2004 |
| Externally published | Yes |
Fingerprint
All Science Journal Classification (ASJC) codes
- Computer Science(all)
- Biochemistry, Genetics and Molecular Biology(all)
- Theoretical Computer Science
Cite this
Using WWW-distribution of words in detecting peculiar web pages. / Hirose, Masayuki; Suzuki, Einoshin.
In: Lecture Notes in Computer Science, Vol. 3245, 2004, p. 355-362.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Using WWW-distribution of words in detecting peculiar web pages
AU - Hirose, Masayuki
AU - Suzuki, Einoshin
PY - 2004
Y1 - 2004
N2 - In this paper, we propose TFIGF, a method which detects peculiar web pages using distribution of words in WWW given a set of keywords. Our TFIGF detects a set of index words which represent a WWW page by estimating their importance in the WWW page and their rareness in WWW. Experiments using both English and Japanese WWW pages clearly show superiority of our approach over a traditional method which employs a limited number of WWW pages in the estimation.
AB - In this paper, we propose TFIGF, a method which detects peculiar web pages using distribution of words in WWW given a set of keywords. Our TFIGF detects a set of index words which represent a WWW page by estimating their importance in the WWW page and their rareness in WWW. Experiments using both English and Japanese WWW pages clearly show superiority of our approach over a traditional method which employs a limited number of WWW pages in the estimation.
UR - http://www.scopus.com/inward/record.url?scp=33751108639&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33751108639&partnerID=8YFLogxK
M3 - Article
AN - SCOPUS:33751108639
VL - 3245
SP - 355
EP - 362
JO - Lecture Notes in Computer Science
JF - Lecture Notes in Computer Science
SN - 0302-9743
ER -