How to get SVM to show predictions?

Hi guys! I am doing a machine learning assignment and need to create an SVM machine learning model that will predict the labels for a variety of rows. I have an excel sheet with rows (tweets) associated with a 0 and 1 label, not fake and fake news. I need to somehow show that SVM can predict the ouput of a tweet as fake or not fake.

I have done BOW and data cleaning, and don’t know if I need to manually remove labels from my excel sheet to allow for predicting? And then how do I show that the SVM predicted the 0 or 1 label for a tweet? Please, any help appreciated! I’ve researched and used tones of nodes but coming to my ends wits. Thank you.

Hi @dilaracet -

Take a look at this example workflow from the KNIME Hub:

It doesn’t use SVM, but you could easily swap out the appropriate nodes with SVM learner / predictor nodes to achieve your result. Give this a shot and if you have questions, feel free to ask here :slight_smile:

1 Like

HI @ScottF Thanks for this. Just while I have you, can you please tell me how I can fix “Execute failed: (“NullPointerException”): null” error with the SVM predictor node? Thanks. :slight_smile:

To troubleshoot I think we need more information here. Would it be possible for you to upload a workflow that reproduces the SVM NPE problem, perhaps with some dummy data?

EDIT: I see now that you’ve opened a separate thread for the NPE issue. Let’s move this discussion there.

1 Like

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.