Regression Using R

Install R and R packages

• 1. Install R from https://cran.r-project.org/bin/windows/base/

• 1. Install RStudio from https://www.rstudio.com/products/rstudio/download/ which workspace can be used to develop R programs

• 1. Install rJava Package by executing the below command in R command window.

Choose a mirror from the popup window to install rJava. It would be better to select USA (KS)

If it takes too much time, click on stop. Open R studio, and then navigate to Packages tab in right side panel and click on install option.

Type rJava and click on install.

• 1. Enter install.packages(‘Rserve’) in R command window. Or else Navigatge to Packages in RStudio and click on install to enter Rserve.

R setup is ready with its packages. We can run R programs in R or in R Studio.

Integrate R With Java using rJava

• 1. Right click on My Computer and select Properties.

  2. In the System window that appears, select Advanced system settings.

  3. System Properties dialog will appear, on this dialog select Environment Variables.

  4. Under System variables section search for variable PATH. Select PATH variable and click on Edit button.

  5. Append C:\R\R-3.2.1\bin\i386;C:\R\R-3.2.1\library\rJava\jri\;C:\R\R-3.2.1\library\rJava\jri\i386;

• 1. Click ok ok ok.

  2. Create a java project in Eclipse. Say, RJRISamples

  3. Right click on project and select BuildPath à Configure Build Path

• 1. Click on Add External Jars in Libraries tab. Select JRI jars from the R installed path.

• 1. "Path to JRI DLL Dir" : This is the directory that contains jri.dll. When you install the rJava package, a folder called 'rjava' is created in the folder that stores all installed libraries in R. if you dont know where that is, open up R and fire the command 'library()'. This will open up another window that shows the path where the libraries are installed. Look for rjava. The line on the top gives the installation path. select the [rjava]/jri folder from the preference page. If your machine is 64 bit select [rjava]/jri/x64

  2. "Path to JRI JARS" : Select the [rjava]/jri folder. Make sure that this folder contains JRI.jar, JRIEngine.jar, REngine.jar. This path is generally same as the one above, unless you want to put the jars in a different location. If any of the above jar is missing then you can download them from http://www.rforge.net/JRI/files/ .

  3. "Path to R DLL Dir" : This specifies the path to R.dll. It should be present in R_Home/bin or R_home/bin/x64 for 64 bit machines.

  4. "Path to JVM DLL dir" : This specifies the path to jvm.dll. It should be present in JRE_HOME/bin/server or JRE_HOME/bin/client.

  5. Click on Apply and then OK. Now you are ready to create your first application.

  6. Create Java application

     • Create New Java Project.

     • Right click on the Java project -> select 'Build Path' -> click 'Add Libraries'.

     • In the popup that comes up, select 'User Library' and click 'Next'.

     • you should see a library called 'JRI_DIR'. Select it and click Finish. You are now ready to write your first class. (If you don't see the JRI_DIR then simply add the jars from [rjava]/jri/ to your classpath.)

Integrate R with java using Rserve:

• 1. Follow Rserve installation steps to install Rserve if not installed previously.

  2. Start Rserve as follows from R

• 1. Create a sample project in Eclipse using Rserve. Right click on Project select Build Path à Configure Build Path. Add the below jars from Rserve installed packages

• 1. Output:

Call RScript in Java

• 1. Create an R sample in RStudio or textpad.

• 1. Create Java code and Run

     1. /**

     2. *

     3. */

     4. package com.tms.rundown.r.rserve.samples;

     6. import org.rosuda.REngine.REXPMismatchException;

     7. import org.rosuda.REngine.Rserve.RConnection;

     8. import org.rosuda.REngine.Rserve.RserveException;

     10. /**

     11. * @author 157330

     12. *

     13. */

     14. public class RScriptSample {

     16. public static void main(String[] args) {

     18. try {

     19. int num1 = 10;

     20. int num2 = 20;

     21. RConnection conn = new RConnection();

     22. conn.eval("source('C:\\\\Rworkspace\\\\RSamples\\\\SumXY.R')");

     24. int sum = conn.eval("addFunc("+num1+", "+num2+")").asInteger();

     25. System.out.println("Sum: "+sum);

     26. conn.close();

     27. } catch (RserveException e) {

     28. // TODO Auto-generated catch block

     29. e.printStackTrace();

     30. } catch (REXPMismatchException e) {

     31. // TODO Auto-generated catch block

     32. e.printStackTrace();

     33. }

     34. }

     35. }

• 1. Output:

Integrate R with Java using rCaller

• 1. Download rcaller jar from http://code.google.com/p/rcaller/downloads/detail?name=RCaller-2.1.1-SNAPSHOT.jar&can=2&q=

  2. Create a java project.

  3. Add rcaller jar to project classpath.

• 1. Create Java code to integrate with R

     1. /**

     2. *

     3. */

     4. package com.tms.rundown.r.rcaller.samples;

     6. import rcaller.RCaller;

     7. import rcaller.RCode;

     9. /**

     10. * @author 157330

     11. *

     12. */

     13. public class RCallerSample {

     15. public static void main(String[] args) {

     16. RCaller rc = new RCaller();

     17. rc.setRscriptExecutable("C:\\R\\R-3.2.1\\bin\\RScript.exe");

     18. double[][] matrix = new double[][]{{6, 4}, {9, 8}};

     20. RCode code = new RCode();

     22. // Passing Java objects to R

     23. code.addDoubleMatrix("x", matrix);

     24. code.addRCode("s <- solve(x)");

     25. rc.setRCode(code);

     27. // Performing Calculations

     28. rc.runAndReturnResult("s");

     30. // Passing R object to Java

     31. double[][] inverse = rc.getParser().getAsDoubleMatrix("s", 2, 2);

     32. System.out.println("Inverse: "+inverse);

     34. }

     36. }

  2. Output: Runtime error.

Integrate R with java using Rengin

• 1. Download renjin-script-engine-0.7.0-RC7.jar from http://nexus.bedatadriven.com/content/groups/public/org/renjin/renjin-script-engine/0.7.0-RC7/

  2. Create a java project and add this jar to classpath.

• 1. Create a java code

package com.tms.rundown.r.renjin.samples;

/**

*

* @author 157330

*

*/

import java.util.*;

import javax.script.*;

public class TryRenjin {

public void test () throws Exception {

// create a script engine manager

ScriptEngineManager manager = new ScriptEngineManager();

// create a Renjin engine:

ScriptEngine engine = manager.getEngineByName("Renjin");

// check if engine has loaded correctly:

if (engine == null) {

throw new RuntimeException ("Renjin Script Engine not found on the classpath.");

}

else {

System.out.println ("Renjin Script Engine initialized!");

}

// run R script coded manually

TestLinearRegression (engine);

// run R script from an external .r script file (located in /src/scripts/.)

TestLinearRegressionFromRFile (engine);

}

private void TestLinearRegression (ScriptEngine engine) {

try {

engine.eval ("df <- data.frame(x=1:10, y=(1:10)+rnorm(n=10))");

engine.eval ("print (df)");

// NOTE: The ScriptEngine won’t print everything to standard out like the

// interactive REPL does, so if you want to output something, you’ll need

// to call the R print() command explicitly.

engine.eval ("print(lm(y ~ x, df))");

}

catch (Exception e) {

System.out.println (e);

}

}

private void TestLinearRegressionFromRFile (ScriptEngine engine) {

try {

System.out.println ("here " + (new Date ()));

engine.eval (new java.io.FileReader ("C:\\Rworkspace\\RSamples\\Script.R"));

}

catch (Exception e) {

System.out.println (e);

}

}

public static void main(String[] args) throws Exception {

TryRenjin renjin = new TryRenjin();

renjin.test();

}

}

• 1. Script.R

     2. df <- data.frame(x=1:10, y=(1:10)+rnorm(n=10))

     3. print(df)

     4. print(lm(y ~ x, df))

  2. Output: <lm à Linear regression in java with R requires jdk 7 to use renjin : java.util.objects>