[SOLVED] Word Net

faridehbagherzafaridehbagherza Member Posts: 22 Contributor II
edited July 2019 in Help
Hi!
I am using Rapid miner 5.3.013. I installed "text mining" and "word Net" extensions. I also downloaded the "word Net" dictionary itself. But I don`t know how to set the parameters of "Open Word Net Dictionary" Operator. I tried everything which came to my mind, But it still says: "Couldn`t open dictionary from directory"

Answers

  • faridehbagherzafaridehbagherza Member Posts: 22 Contributor II
    Here is a little more explanation, I took it from a previous thread which was not solved:
    "To use Wordnet with the Word Vector Tool, you need a working installation of Wordnet 2.1.  Also, you need a configuration file for JWNL. An example configuration file can be found in the sample directory. Usually it should be sufficient to set the correct path to your Wordnet dictionary directory (setting the parameter ‘dictionary_path’). For more information on configuring the JWNL please refer to their homepage.  Currently, Wordnet is supported for the use in the stemmer step, thus to reduce a word to some base form. The corresponding classes are called ‘WordNetHypernymStemmer’ and WordNetSynonymStemmer’."

    BUT:
    (1) Where do I put the JWNL files?
    (2) Where do I put the Wordnet files?
    (3) Which file should I import in "Open Word Net Dictionary" Operator?
  • Nils_WoehlerNils_Woehler Member Posts: 463 Maven
    Hi,

    you have to download the WordNet 3.0 database files (http://wordnetcode.princeton.edu/3.0/WNdb-3.0.tar.gz). The folder has to be extracted to somewhere and the "Open WordNet dictionary" operators has to point to the folder that you have extracted.

    Best,
    Nils
  • faridehbagherzafaridehbagherza Member Posts: 22 Contributor II
    Thanks Nils
    It worked!
    I am so delighted.

    Best
    Farideh
  • calangcalang Member Posts: 6 Contributor II
    Farideh, Nils,

    I have the same initial questions Farideh had, especially

        Where do I put the JWNL files?

    Do you have some indication about how to do that?

    My model is working "fine", having the Open WordNet Dictionaty operator pointing to the directory where I have the text files for the WN dictionary, but no word or token is recognized from my input document.  It seems no word is being read from the dictionary.

    Any suggestions, please?
    Thanks.
  • evelyne_tran_1evelyne_tran_1 Member Posts: 9 Contributor I

    Hi @Nils_Woehler,

     

    Compared to Farideh, I didn't manage to open the WordNet Dictionary even if I followed your steps... 

     

    You will find attached to this message the error I am facing...

     

    Can you help please?

     

    Thanks !

  • Thomas_OttThomas_Ott RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,761 Unicorn

     Wordnet can be pretty finicky to get right.  There are couple of silly gotchas that I ran into over time. One is making sure the order of execution is right. You might want to check that the Open Wordnet operator loads first before the everything else inside the operator loads. 

     

    Second is making sure the path is correct (I believe it's the /dict folder) and that the resource type is set as directory. 

     

    With regard to extracting all the files. Since I run Windows, I extract the entire tar.gz file into a directory and then path from there. 

Sign In or Register to comment.