This file contains the information needed to run java programs.

Each algorithm project has precompiled jar file inside bin folder, this jar file can be used right away without 
the need to compile the code. Each algorithm project has csv files needed to run the java program. Those files 
can be changed and duplicated if needed.

Spark PkNN:
To run this algorithm copy the jar file to the cluster with Spark installed, copy test and train csv files to HDFS
and run the command below:

	java -jar -Dspark.executor.memory=979m SparkPkNN.jar 5 < path to test data in HDFS > 4

Here 5 is a number of neighbors and 4 the number of workers to use

Hadoop PkNN:
To run this algorithm copy the jar file to the cluster with Hadoop installed, copy test and train csv files to HDFS
and run the command below:

	hadoop jar HadoopPkNN.jar 5 < path to test data in HDFS >

Here 5 is a number of neighbors

Spark PnBm:
To run this algorithm copy the jar file to the cluster with Spark installed, copy test and train csv files to HDFS
and run the command below:

	java -jar -Dspark.executor.memory=979m SparkPnBm.jar < path to test data in HDFS > 4

Here 4 is the number of workers to use. In addition there is PnBm train algorithm project that can be used to generate train data file

Hadoop PkNN:
To run this algorithm copy the jar file to the cluster with Hadoop installed, copy test and train csv files to HDFS
and run the command below:

	hadoop jar HadoopPnBm.jar < path to test data in HDFS >

Spark Clara:
To run this algorithm copy the jar file to the cluster with Spark installed, copy test and train csv files to HDFS
and run the command below:

	java -jar -Dspark.executor.memory=979m SparkClara.jar 512 20 100 2 < path to test data in HDFS >

Here 512 is a sample size, 20 number of medoids, 100 number of experiments and 4 the number of workers to use

Hadoop Clara:
To run this algorithm copy the jar file to the cluster with Hadoop installed, copy test and train csv files to HDFS
and run the command below:

	hadoop jar SparkClara.jar 512 20 100 2 < path to test data in HDFS > 4

Here 512 is a sample size, 20 number of medoids, 100 number of experiments and 4 the number of reduce tasks to use

Additionally there is test file generator project that can be used to generated test data for each algorithm.