/cluster-mapred

simple cluster code based on map reduce

Primary LanguageJava

Clustering Example

  • In many clustering applications the first step is to find the equivalent entities and build the clusters from these
    hear we can exploit the fact that map-reduce aggregates the identical keys, by selecting the key as the match entity
    we can create initial seed clusters, then we an apply a metric to aggregate the remaining data to these centers

    100% equivalent node1 node2 .....
    90% equivalent subnode 1a subnode 1a ............
    repeat

  • To run code

  • 1 run com.eduonix.hadoop.partone.EntityAnalysisMRJob to produce the traning data, set
    public static final boolean runOnCluster = false for local testing on Linux
    public static final boolean runOnCluster = true for Hadoop
  • 2 run com.eduonix.hadoop.partone.etl.EntityAnalysisETL