Name Stemming

I did a quick search for

larry summers firing line site:pbs.org

But it did not seem like Mojeek recognized larry as a stem for Lawrence

I did not find the correct link anywhere in the results while searching for larry

Only searching with Lawrence showed the correct result. It was even ranked first.

Question

Is name stemming a feature that Mojeek is planning on adding?

References

Stemming - Wikipedia

Mike

2 Likes

Hey Mike,

This is for sure an understemming issue and it’s something we’re aware of and will be looking at in the coming weeks - and not just for names - with a view to improving how it is handled.

Thanks for the example as it has been added to a few that we’ve had in recently which show where Mojeek is falling down.

Also, I wasn’t aware that Firing Line had been rebooted. Courtesy of the Hoover Institution Archive’s YouTube Channel (sorry for pointing at a G**gle property) I’ve watched a load of the original, I will definitely check out this.

3 Likes

“Mauritian” was not recognized as a stem for “Mauritius” while using my Focus Engine.

Image

{"Countries":"i=www.cia.gov"}

I expected to see this page.

I’m using the new algorithm.

Image

Thanks for pointing this one out @mike, interestingly enough those two words are not stemmed at all. Can’t really say how you’d solve this one, but it has gone into the “things to think about” pile.

1 Like