Match short phrases with long text input

Hi Alteryx Community,

I'm currently working on some matching and labelling task that requires the technique of fuzzy match.

I have read through a few posts about fuzzy match on the forum, and seems that majority of them are about short text to short text match. I would like to seek for some suggestions about how to fuzzy match between short text and long paragraph. Please refer to the sample data file attached. Basically, I'm aiming to match the keyword column to the Raw text column and label each record of the raw text with the corresponding keyword. For some raw text, it may contain a certain part of the keyword phrase, but it may also not be the correct match to the keyword phrase due to the lack of exact location name.

Thanks!

Sample details.xlsx

Workflow

Text Mining

Accepted answers

ChrisTX

Maybe try a 2-phase approach. First the exact match, then the Fuzzy match. You may be disappointed in the output from the Fuzzy match. It's never perfect, as expected.

Check out the video training for Fuzzy Match on this page: Videos - Alteryx Community

Chris

All comments

ChrisTX

Your example data looks like you're matching on exact words. Try the Text to Column tool, but split to rows. See attached workflow.

Screenshot 2024-08-26 053320.png

Fuzzy match may not be needed in this case. I only use Fuzzy Match when there's no other choice.

Chris

Match short phrases with long text input.yxmd

WuQi

Hi Chris,

Thanks for the reply! The keyword column in the sample represents location information and the raw text column is from customer's input. Besides exact match, I'm also considering the cases for misspellings, possible short form of the location names, etc. That's why I'm seeking for suggestions on fuzzy match.

Thanks

ChrisTX

Maybe try a 2-phase approach. First the exact match, then the Fuzzy match. You may be disappointed in the output from the Fuzzy match. It's never perfect, as expected.

Check out the video training for Fuzzy Match on this page: Videos - Alteryx Community

Chris

Quick Links

This months top contributors

atcodedog05 19458

Qiu 15866

binu_acs 15708

MarqueeCrew 13708

apathetichell 13703