K-means clustering is a very popular algorithm used for clustering data. This extension uses the PySpark MLlib implementation of this algorithm. In order to run K-means clustering, you need to specify the number of clusters you want. This can be determined using domain knowledge about your dataset or through trial and error of evaluating different cluster parameters.
Learn more about this implementation from the MLlib Documentation
- SPSS Modeler v18.0 or later
- Python 2.7 Anaconda Distribution
More information here: IBM Predictive Extensions
If using v18.0 of SPSS Modeler, navigate to the options.cfg file (Windows default path: C:\Program Files\IBM\SPSS\Modeler\18.0\config). Open this file in a text editor and paste the following text at the bottom of the document:
eas_pyspark_python_path, "C:/Users/IBM_ADMIN/Anaconda/python.exe"
- The italicized path should be replaced with the path to your python.exe from your Anaconda installation.
- Go to the Extension menu in Modeler and click "Extension Hub"
- In the search bar, type the name of this extension and press enter
- Check the box next to "Get extension" and click OK at the bottom of the screen
- The extension will install and a pop-up will show what palette it was installed to
- Save the .mpe file to your computer
- In Modeler, click the Extensions menu, then click Install Local Extension Bundle
- Navigate to where the .mpe was saved and click open
- The extension will install and a pop-up will show what palette it was installed
- Nial McCarrol - (www.mccaroll.net)
- Greg Filla (gdfilla)