Introduction

TECHMARK-Maximizing-ROI

Weka Installation on Mac

Download Weka disk image for OS X and install it.

Install libsvm

By default Weka doesn’t has libsvm library. In order to use SVM classification algorithm, we have to install the libsvm jar package manually.

Install libsvm.jar step by step:

• Download libsvm, and save to a local directory.
• Unzip the libsvm-x.xx.zip file.
• Copy the jar file java/libsvm.jar to the directory /Applications/weka-3-6-13-oracle-jvm.app/Contents/Java/Classes. This directory is in the classpath of Weka
• Go to directory /Applications/weka-3-6-13-oracle-jvm.app/Contents/Java/Classes, extract the libsvm.jar by the command jar xf libsvm.jar

Create Data Set

Training Data Set

For Rehmertz 2, create training data file rehmertz-2-training.arff.

%
% Rehmertz Training Data
% Reference to the table-1 "实验室中的模拟试验 - 雷墨磁2"
%
@relation Rehmertz

@attribute torque-diff numeric
@attribute resistance-diff numeric
@attribute cost numeric
@attribute result {pass, fail}

@data
1.0,0.4,0.18,fail
0.7,0.9,1.15,pass
4.0,4.0,4.00,fail
4.0,4.0,6.00,pass

For Rehmertz 1 and Rehmertz 3, create training data file rehmertz-1-3-training.arff.

%
% Rehmertz Training Data
% Reference to the table-2 "实验室中的模拟试验 - 雷墨磁1和雷墨磁3"
%
@relation Rehmertz

@attribute torque-diff numeric
@attribute resistance-diff numeric
@attribute cost numeric
@attribute result {pass, fail}

@data
1.0,0.4,0.4,fail
0.7,0.9,1.1,pass
4.0,4.0,4.00,fail
4.0,4.0,6.00,pass
0.7,0.8,0.75,fail
1.4,1.0,1.25,pass

Test Data Set

Create testing data file rehmertz-test.arff.

%
% Rehmertz Testing Data
%
@relation Rehmertz

@attribute torque-diff numeric
@attribute resistance-diff numeric
@attribute cost numeric
@attribute result {pass, fail}

@data
3,3,0.2,pass
3,3,0.4,pass
3,3,0.6,pass
3,3,0.8,pass
3,3,1.0,pass
3,3,1.2,pass
3,3,1.4,pass
3,3,1.6,pass
3,3,1.8,pass
3,3,2.0,pass
3,3,2.2,pass
... ...
3,3,8.0,pass

Detail Steps

Load traing data set

• Open Weka GUI and click the “Explorer”

  

• In the “Weka Explorer” Window, select “Preprocess” tab, and click the Open File button and select the file rehmertz-2-training.arff. The the training data for Rehmertz 2 is loaded.

  

Select Classification Algorithm

• Select the Classify tab in the “Weka Explorer” Window, and click choose button in Classifier area

  

Select the classification algorithm logistic

Supply test data file

• Under the same Classify tab, in the Test options area, select Supplied test set and click the Set-> Open file... to the load the file rehmertz-test.arff.

Select Classification Options

• Click More options in the Test options area. In the Classifier evaluation options window, select Output predictions option, and set Output additional attributes to 3.

Run Classfication and Check Result

• Confirm the (Nom) result is selected as target. Then click start

• The execution result is shown