READ ME.txt

K-Means Clustering Algorithm Implementation Using MapReduce 
by Vishal Doshi
The MapReduce application is packed in 659391383_ASSIGN2.zip. Extract and import in Eclipse to view the source code.
Class and files:
KMeansFileGenerator.java: used to generate data points at random and random centroids are chosen from data generated.  User define datapoint generated with 4 datapoint chosen to be centroids at random.
KMeansHadoop: Driver class for the MapReduce Application
KMeansMapper: Mapper maps datapoints to closest cluster
KMeansReducer: Reducer recalculate the centroids and writes them
KMeansPartition: Partitioner assigns reducer according to the cluster id.
Instructions to run: 
Step 1: Copy KMeansFileGenerator.java to desktop. Compile and run.
•	javac KMeansFileGenerator.java
•	java KMeansFileGenerator.
Step 2: Generate datapoints.txt and centroid_1.txt and move it on Desktop
Step 3: Create ‘kmeansInput’ and ‘centroid’ directory in HDFS
•	./bin/hadoop fs -mkdir kmeansInput
•	./bin/hadoop fs -mkdir centroid
Step 4: Move the files from to Desktop to HDFS
•	./bin/hadoop fs -put ~/Desktop/datapoints.txt kmeansInput
•	./bin/hadoop fs -put ~/Desktop/centroid_1.txt centroid
Step 5: Export the jar file to desktop and Run .jar file
•	./bin/hadoop jar ~/Desktop/kmeanshadoop.jar KMeansHadoop kmeansInput kmeansOutput/out centroid
[3 parameters - 1: input directory, 2: output directory, 3: centroid file location]