comparing text using machine learning

Hi guys :slight_smile:
I thinking is there any option to compare two text, which containing similar sign or words. e.g.

column1 column2

John,20-000,sunny john,2000,Sun…ny

some of the sign can be distort.

Hi @pjacob

take a look at the String Similarity Node .
gr. Hans

6 Likes

I would start with Levenshtein distance.

6 Likes

@HansS maybe you know. I have two tables, table 1 contain “Text input 1”, and second table contain “Text input 2”. I want to connect second table to “Table Row To Variable Loop Start” Node and then connect to String Similarity but unforutnatelly give me in each row similarity 1.0 (I used Levenshtein)?

Hey :smiley: I want to check if rows in table are similar to rows in table below. I create something like this but I get only “1.0”
KNIME_project.knwf (12.6 KB)

You’re currently comparing the same column with each other which naturally results in 1.0 :slight_smile:

image

You’ll probably simply need to properly configure the node or restructure your input so that it consists of two columns which you want to compare with each other.

Without knowing your task, you can probably do that with one of the “Joiner” nodes:

image

7 Likes

Hi I cannot find the string similarity node anywhere, I could not find it in Node pit either, is it availble somewhere else?

If you have KNIME 4.2 this has to work for you

1 Like

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