Replies to This Discussion

Permalink Reply by Larry on May 4, 2011 at 6:48am

R has packages for creating decision trees.  The most notable one is rpart.  For a variety of add-on packages that R can use for data mining, including decision trees, look at the Machine Learning task views page.

 

http://cran.r-project.org/web/views/MachineLearning.html

 

 

• ▶ Reply

Permalink Reply by Richard on May 4, 2011 at 10:49am

Check C4.5. 

 

C4.5 is a software extension of the basic ID3 algorithm designed by Quinlan to address the following issues not dealt with by ID3:

• Avoiding overfitting the data
  • Determining how deeply to grow a decision tree.
• Reduced error pruning.
• Rule post-pruning.
• Handling continuous attributes.
  • e.g., temperature
• Choosing an appropriate attribute selection measure.
• Handling training data with missing attribute values.
• Handling attributes with differing costs.
• Improving computational efficiency.

It is installed for use on Grendel (grendel.icd.uregina.ca), but it may be set up on a local machine as follows:

C4.5 Release 8 Installation Instructions for UNIX

1. Download the C4.5 source code.
2. Decompress the archive:
   1. Type "tar xvzf c4.5r8.tar" (not universally supported), or, alternatively,
   2. Type "gunzip c4.5r8.tar.gz" to decompress the gzip archive, and then
      Type "tar xvf c4.5r8.tar" to decompress the tar archive.
3. Change to ./R8/Src
4. Type "make all" to compile the executables.
5. Put the executables into a "bin" subdirectory and include it in the path for command-line usage.

Manual Pages

• c4.5: using the c4.5 decision tree generator.
• verbose c4.5: interpreting output generated by c4.5.

• c4.5rules: using the c4.5 rule generator.
• verbose c4.5rules: interpreting output generated by c4.5rules.

• consult: uses a decision tree to classify items.
• consultr: uses a rule set to classify items.

Examples

Click on the links below for examples of C4.5 usage:

• Example 1 - Golf
  • A simple, detailed example of how C4.5 and C4.5rules work.
• Example 2 - Sunburn
  • The sunburn example revisited.
• Example 3 - Homonyms
  • Advanced usage of, and a practical application of, C4.5 and C4.5rules.

Source: http://www2.cs.uregina.ca/~dbd/cs831/notes/ml/dtrees/c4.5/tutorial....

• ▶ Reply

Permalink Reply by Abhinav Jain on May 5, 2011 at 10:26am

You can also use WEKA. It is open source data mining software with wide variety of machine learning algorithms for data mining tasks. It's easily installable and GUI based easy to use.

Google it or here is the link:

http://www.cs.waikato.ac.nz/ml/weka/

 

I'm using WEKA for my data mining course work, so let me know in case you need any assistance.

 

• ▶ Reply

Permalink Reply by Yashu Gupta on May 5, 2011 at 10:34pm

thanks Abhinav ....

I will take a look and get back to you in case of any clarifications..

• ▶ Reply

Permalink Reply by Yashu Gupta on May 5, 2011 at 11:11pm

Hi Abhinav,

 

I am able to successfully download and install WEKA.

Now, I have CSV file having variables and there corresponding data. I want to build a decision tree in which in which one variable will be performance variable and other will be independent variables.

Could you please guide me steps in creating this kind of decision tree in WEKA....?

• ▶ Reply

Permalink Reply by sana on May 5, 2011 at 11:36am

weka is the best, from learning perspective. The best thing is, it operates on text files, so from learning perspective you just need the sample file and you are good to go.,

• ▶ Reply

Permalink Reply by Yashu Gupta on May 10, 2011 at 6:17am

Hi Abhinav,

 

I am able to convert the CSV into ARFF file.

Could you please guide me the steps that i need to follow to create decision tree. My file has 4 variables in which 3 are categorical and one is numeric (dependent variable).

 

I just want to split the 3 categorical variables based on this fourth numeric dependent variable

 

Please advice. Thanks in advance.

 

Regards,

Yashu

• ▶ Reply

Permalink Reply by Bhupendrasinh Thakre on July 27, 2012 at 12:10pm

Hi Yashu,

Actually you dont have to convert csv file at all.

When selecting the file to load just select "csv" and program will pop only csv files.

So you have 3 independent var and 1 dependent which is numeric. Regression Tree might help.

Best,

Bhupendrasinh Thakre

• ▶ Reply