j48+knn+group model

[Deleted User]
[Deleted User] New Altair Community Member
edited November 2024 in Community Q&A
Hi 
would you please tell me how can I show the graph for KNN, J48 and group model?
thank you
Tagged:

Best Answers

  • rfuentealba
    rfuentealba New Altair Community Member
    Answer ✓
    Hi @mbs,

    Can confirm:
    • Weka W-J48 and Decision Tree are based in the C4.5 algorithm, so @sgenzer is right and I haven't found any differences between both. If you need a visualization of the model, you are better off using the C4.5 Decision Tree there.
    • k-NN (Nearest-Neighbor) models (there are many, depending on what measure you use) cannot be put in a chart the same way a Decision Tree can, but if you want to visualize the results, a scatter plot may help.
    • Group models, the same, you should use scatter plots to visualize the data, but there is no graphical representation of the model itself.
    All the best,

    Rodrigo.
  • varunm1
    varunm1 New Altair Community Member
    Answer ✓
    Hello,

    The only model visualization method I can see is by using Ungroup models operator that will show individual models in a group model. 

Answers

  • [Deleted User]
    [Deleted User] New Altair Community Member
    @sgenzer
    would you please help me?  :)
  • sgenzer
    sgenzer
    Altair Employee
    @mbs the W-J48 model, as with all the Weka models, are very dated and do not show very well. I would strongly recommend using the RapidMiner Decision Tree operator rather than the W-J48. You can see here that RapidMiner's Decision Tree model is based on the same C4.5 algorithm as the W-J48. And of course you can inspect the code via GitHub (see that link above) if you need transparency for research purposes, etc...

    As for k-NN, I do not know of any way to visualize the models. 

    Scott
  • rfuentealba
    rfuentealba New Altair Community Member
    Answer ✓
    Hi @mbs,

    Can confirm:
    • Weka W-J48 and Decision Tree are based in the C4.5 algorithm, so @sgenzer is right and I haven't found any differences between both. If you need a visualization of the model, you are better off using the C4.5 Decision Tree there.
    • k-NN (Nearest-Neighbor) models (there are many, depending on what measure you use) cannot be put in a chart the same way a Decision Tree can, but if you want to visualize the results, a scatter plot may help.
    • Group models, the same, you should use scatter plots to visualize the data, but there is no graphical representation of the model itself.
    All the best,

    Rodrigo.
  • varunm1
    varunm1 New Altair Community Member
    Answer ✓
    Hello,

    The only model visualization method I can see is by using Ungroup models operator that will show individual models in a group model. 
  • [Deleted User]
    [Deleted User] New Altair Community Member
    hi every body
    thank you for your help i will try all of them :)
    mbs