The JMP® Scripting Language (JSL) script, and data table were presented at the April 2, 2015 Mid-Atlantic JMP User Group (MAJUG) meeting. They reviewed the ways JMP® can apply the techniques of text mining. This presentation was based on a E-poster Joshua Klick and I, MAJUG co-chairs, presented at the Discovery Summit 2014 conference (“Text Mining Feedback Comments from JMP® Users Group Meeting Participants”, https://community.jmp.com/docs/DOC-6748).
The E-poster showed how JMP® and R integration transformed, unstructured, free-text comments from respondents to MAJUG meeting feedback surveys. These comments were helpful to MAJUG leadership in planning meetings that met their needs. Many visitors to our E-poster wanted to know how the fundamental text mining tasks (available in SAS® Text Miner, SAS/IML, and R packages) could be done using JMP alone.
JMP® version 11 introduced the Free Text Command in the Analyze > Consumer Research > Categorical Platform under the “Multiple” tab. This tool restricted users to just produce word frequency counts and create indicator columns of the words that appeared in free-text comment columns. For more Extensive text mining, users would need to use other JSL scripts, functions, and tools.
The text mining techniques covered in the data table example and JSL script include: selecting and removing “stop” words; using the RECODE command to “stem” terms; forming Term-Document-Matrices (TDMs) using Julian Parris' “Word Counts to Columns” script, (Word Counts to Columns); applying the singular value decomposition (SVD) function to identify the underlying dimensions that accounts for most of the information made by respondents and text; and clustering word groups to convey similar topics, themes, or synonymous terms. With the JMP® tools presented, users will be able to apply the methods to mine their own textual data.
