---------------------------
Affinity Propogation on Hadoop Cluster.
---------------------------
This is a java-distributive implementation of Affinity Propagation.
Any questions or comments about this code should be sent to songxiaoli2012@gmail.com.
------------------------------------------------------------------------
TABLE OF CONTENTS
A. PREREQUISITE.
B. DATA INPUT FORMAT.
C. RESULT.
------------------------------------------------------------------------
A. PREREQUISITE.
The Hadoop cluster needs to be installed before running the jar files.
------------------------------------------------------------------------
B. DATA INPUT FORMAT.
The data is a file where each line is of the form:
[M] [term_1] [term_2] ... [term_N]
* [M] is the document id.
* [term_i] is the i-th term in document M.
------------------------------------------------------------------------
C. RESULT.
The result is in the folder <working directory>/Cluster. It gives each document (given as document id) to a cluster (represented by the cluster center given as document id.)
评论1