Main.ThomasHofmannCiteSeer
---
Data: Citeseer
Task:
a) Deduplication (removal of duplicates) of citations (i.e. matching same citations, that are differently spelled
b) Cluster (with overlapping clusters) authors according to their interest. Idea: cluster the links - non-overlapping
Method: generative model. Given prob(group), prob(entity | group) : merge clusters with lots of common entities
Motto: "Clusters are similar if linkssets are similar"
Rating: 3/5
---
Only b) seems to be interesting for me. But this group formation algorithm can be applied to communities in general.
Пожалуйста, войдите в систему, чтобы принять участие в дискуссии (добавить собственные рецензию, или комментарий)
Цитировать эту публикацию
%0 Conference Paper
%1 citeulike:344957
%A Bhattacharya, Indrajit
%A Getoor, Lise
%B KDD Workshop on Link Analysis and Group Detection
%C Seattle, WA
%D 2004
%K group link
%T Deduplication and Group Detection using Links
%U http://www.cs.umd.edu/~getoor/Publications/linkKDD04.pdf
@inproceedings{citeulike:344957,
added-at = {2007-08-15T12:04:46.000+0200},
address = {Seattle, WA},
author = {Bhattacharya, Indrajit and Getoor, Lise},
biburl = {https://www.bibsonomy.org/bibtex/24ae90bacefb2df99d3003257d6eca8ca/wnpxrz},
booktitle = {KDD Workshop on Link Analysis and Group Detection},
citeulike-article-id = {344957},
comment = {Main.ThomasHofmannCiteSeer
---
Data: Citeseer
Task:
a) Deduplication (removal of duplicates) of citations (i.e. matching same citations, that are differently spelled
b) Cluster (with overlapping clusters) authors according to their interest. Idea: cluster the links - non-overlapping
Method: generative model. Given prob(group), prob(entity | group) : merge clusters with lots of common entities
Motto: "Clusters are similar if linkssets are similar"
Rating: 3/5
---
Only b) seems to be interesting for me. But this group formation algorithm can be applied to communities in general.},
interhash = {edccfa6945a6525993cc0bdbffbf731a},
intrahash = {4ae90bacefb2df99d3003257d6eca8ca},
keywords = {group link},
month = {August},
priority = {0},
timestamp = {2007-08-15T12:04:46.000+0200},
title = {Deduplication and Group Detection using Links},
url = {http://www.cs.umd.edu/~getoor/Publications/linkKDD04.pdf},
year = 2004
}