Homec4science
Diffusion invenio-infoscience (master) Browse

R3600/invenio/legacy/bibrank/doc/hacking/bibrank-pagerank-t.webdoc
dd44fdec50bbmaster

Last Change
Raw File

bibrank-pagerank-t.webdoc
View Options

## -*- mode: html; coding: utf-8; -*-
## This file is part of Invenio.
## Copyright (C) 2010, 2011 CERN.
##
## Invenio is free software; you can redistribute it and/or
## modify it under the terms of the GNU General Public License as
## published by the Free Software Foundation; either version 2 of the
## License, or (at your option) any later version.
##
## Invenio is distributed in the hope that it will be useful, but
## WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
## General Public License for more details.
##
## You should have received a copy of the GNU General Public License
## along with Invenio; if not, write to the Free Software Foundation, Inc.,
## 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA.
<!-- WebDoc-Page-Title: Time-dependent Link-based Ranking Method -->
<!-- WebDoc-Page-Navtrail: <a class="navtrail" href="<CFG_SITE_URL>/help/hacking">Hacking Invenio</a> &gt; <a class="navtrail" href="bibrank-internals">BibRank Internals</a> -->
<!-- WebDoc-Page-Revision: $Id$ -->
<pre>
<blockquote>
This ranking method combines the link-based ranking with the time-dependent
citation counts. This ranking method is not very well suited for data
collections that allow cycles, since for strong time decay parameters,
it tends to overestimate the value of recent publications.
<b>Algorithm:</b>
1. Read the citation data from the database or from a file
that has the following format: x[tab]y where the paper with
recid x cites the paper with recid y. The citation data is
stored in a map key:value, where the 'key' is a record id
and the 'value' is the list of publications that cite
publication 'key'.
2. Read the publication dates for each paper from the
database or from a file that has the following format: x[tab]y,
where x is a recid and y is the publication year. There are
several possibilities for retrieving the dates from the database:
i) using the 260__c MARC tag that contains only the publication
year (this is the option that we are using);
ii) using the 269__c MARC tag that contains the complete
publication date.
For the papers that do not have a publication date, we consider
the date of insertion in the database (961__x tag). If neither
the publication date nor the insertion date are available,
we consider an average date (computed with the existing dates).
3. Read the convergence threshold, check_point and damping_factor
parameters from the configuration file. (parameters specific
for the link-based ranking)
4. Read the time_decay factor from the configuration file.
5. Remove the simple loops (loops caused by x cites y and
y cites x, and loops caused by x published in year t1 cites
y published in year t2, but t1 < t2).
6. Calculate the weight of each citation based on its age
and time_decay.
7. Calculate the final rank for each publication.
8. Write the ranks to the database and to a file. The name of
the file and the number of ranks that should be outputted can
be set into the configuration file. If the name of the file
is not set, then the ranks are only written in the database.
</blockquote>
</pre>
c4science · Help