For starters, you would need to find or create text files for sentiment, negation, and emphasis terms (see the article I cited for details). If you set them according to the Utilities tab, then the procedure will use them to work with Hebrew text. The procedure will not "understand" Hebrew text, but it might work well enough for your purpose. It really depends on what you want to do with the text, so more details on that would be helpful. It would also help to have a list of stopwords (words like a, and, the). The procedure doesn't currently have a direct way to add those, but I could add that, or it might be added to the language-specific parser if one is found.
Beyond that, there might be some code available that would improve the text processing and could be adapted into the TEXT ANALYSIS command, but that would depend on what is found.
--
Original Message:
Sent: 7/25/2023 7:47:00 AM
From: Meni Berger
Subject: RE: Text Analysis Extension- Hebrew Support
So, let me see if I got you right, this extension is not Hebrew-ready "out-of-the-box" although it might be.
You might try adding some term files (*.tsv?) and also the underlying functionality, so does this extension will support Hebrew text analysis?
I and the entire SPSS Israel community will be forever grateful if such a function is readily available in SPSS Stats. It will also give the product a competitive edge in the locale (pun intended) market.
If you need my assistance in such an endeavor, I'll be glad to provide it!
Original Message:
Sent: 7/23/2023 9:58:00 AM
From: Jon Peck
Subject: RE: Text Analysis Extension- Hebrew Support
This extension does not have any built-in support for languages other than English except for the spell checker and some synonym dictionaries. The synonym languages for search do include Hebrew. I was able to add some general support for German, although it does not really handle the complexities of German grammar - particularly the way German verbs sometimes split up.
However, the extension does provide a way to add translated vocabulary for supplemental sentiment scores, negation terms, and emphasis terms. You can see what these are for English in the article, Analyzing Survey Text.pdf, that is installed with the extension. The Utilities tab lets you create datasets that show the built-in terms and scores and lets you supply your own language data for these items.
I found a source of sentiment terms with a little Googling here
There are probably files of these types of terms that you could find online. If you find a set of terms and it seems to work well, I can work with you to add this support into the extension.
--
Original Message:
Sent: 7/23/2023 9:26:00 AM
From: Meni Berger
Subject: Text Analysis Extension- Hebrew Support
Hello, dear friends.
I have a question regarding Text Analysis Extension for @Jon Peck- can this extension handle Hebrew text? if so, which functions will be available? are there any preliminary packages to be installed?
thanks!
------------------------------
Meni Berger
------------------------------