[MAHOUT-933] Implement mapreduce version of ClusterIterator

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.6
• Fix Version/s: 0.7
• Component/s: classic
• Assignee: Jeff Eastman
• Reporter: Paritosh Ranjan
• Created: 20/Dec/11 04:32
• Updated: 31/Jan/24 22:14
• Resolved: 09/May/12 22:05

Description

Right now, ClusterIterator consumes vectors only from in-memory and sequential hdfs. A mapreduce version to consume vectors needs to be implemented.

Issue Links

• is part of: MAHOUT-929 Refactor Clustering (Vector Classification) into a Separate Postprocess with Outlier Pruning

Sub-Tasks

• Convert K-means buildClusters to use new ClusterIterator: Closed
• Convert fuzzy-K-means buildClusters to use new ClusterIterator: Closed
• Convert Dirichlet buildClusters to use new ClusterIterator: Closed
• Convert Canopy, MeanShift, K-means, Dirichlet, Fuzzy KMeans and Other Tools to emit ClusterWritable: Closed