homepage Welcome to WebmasterWorld Guest from 54.205.144.54
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member
Visit PubCon.com
Home / Forums Index / Google / Google News Archive
Forum Library, Charter, Moderator: open

Google News Archive Forum

This 52 message thread spans 2 pages: 52 ( [1] 2 > >     
Singular and plural keywords, Google Stemming
sandalwood




msg:206934
 8:53 am on Nov 27, 2003 (gmt 0)

I don't know if this dance is over, but i am sure G has changed so many things in it's algo.

I have just noticed something:

I did a search with kw(in singular) and inside the first page results i noticed that there were titles with kw(in plural) marked bold.
keyword: single holidays results: single + singles holidays
Has G the ability of understanding the singular and the plural of the kws?

It's the first time i notice that in G.
Has anybody noticed something similar?

 

Just Guessing




msg:206935
 9:49 am on Nov 27, 2003 (gmt 0)

Yes this is coming and going in the results at the moment - now you see it, now you don't.

See thread [webmasterworld.com ]

It's not just plurals, it's word stemming: search for widget and you get

    widgets
    widgeting
    widgetal
    etc - depending on the word

The exact spelling probably still carries more weight in the ranking.

Just Guessing




msg:206936
 10:08 am on Nov 27, 2003 (gmt 0)

Just noticed that Google changed their Search Tips today to say:

Google now uses stemming technology. Thus, when appropriate, it will search not only for your search terms, but also for words that are similar to some or all of those terms. If you search for "pet lemur dietary needs", Google will also search for "pet lemur diet needs", and other related variations of your terms. Any variants of your terms that were searched for will be highlighted in the snippet of text accompanying each result.

Yesterday the Search Tips said they didn't do it because you got more accurate results without it!

sandalwood




msg:206937
 10:13 am on Nov 27, 2003 (gmt 0)

I think that's not correct!

Google can't know what every user has in his mind...
(Maybe G wants to change what's in every user mind:) )

The only use of this new G feature would be in cases that are not many results of a search Term.

Pierre2003




msg:206938
 11:00 am on Nov 27, 2003 (gmt 0)

Perhaps that is a part of the reason for the last crazy update?

Marcia




msg:206939
 11:16 am on Nov 27, 2003 (gmt 0)

Here it is in their help pages, I guess just_guessing wasn't guessing about this.

[google.com...]

>>Google can't know what every user has in his mind...

It actually can make for more relevant search for users. What if someone is looking for a widget and a page that's got just the right widgets just doesn't have the singular, just the plural on the page. It may still be the best result for them.

[edited by: Marcia at 11:18 am (utc) on Nov. 27, 2003]

sandalwood




msg:206940
 11:18 am on Nov 27, 2003 (gmt 0)

A part, yes, could be :)

Just Guessing




msg:206941
 11:48 am on Nov 27, 2003 (gmt 0)

Fact: Google has a dictionary that it uses for word stemming.

Fact: Google Adwords has a dictionary/database it is using for Expanded Broad Matching. This database includes both word stemming variations and different related keywords.

Theory: Google is using the same database as Adwords for word stemming.

Tenuous Theory: Google is doing some Expanded Broad Matching in the normal SERPS using the Adwords database. Obviously this will only affect keywords or phrases which have expanded broad matches in the database.

Conspiracy Theory: Google is using a different algorithm with keyword combinations with high bids in the Adwords database. A threshold of about $1 for top position fits with the results I have seen, but I don't want to believe the evidence!

sandalwood




msg:206942
 12:01 pm on Nov 27, 2003 (gmt 0)

I guess you are right most of it.
By the time dictionary has to do with the results then G would also pay some attention on the grammar too as there are many words that can be singular, plurar or just a verb.

But Web doesn't contain just English.
What about my language? Greek. German,Italian,Spanish etc..?

G is gonna make a dictionary of all these languages too? I don't think so....

Pierre2003




msg:206943
 12:33 pm on Nov 27, 2003 (gmt 0)

I disagree.

The adwords system is quite different than the search engine. While the adwords system tries to provide:
A. better results for users
B. Drive more traffic to advertisers
C. Drive better traffic (at least they say so)

The search engine only wants one:

PROVIDE BETTER AND MORE RLEVANT RESULTS FOR THE USERS

the advertisers/site owners are not in G's scope regarding the regular search. Only the users are.

True, that broad match also helps users... But I find it hard to believe...

Just Guessing




msg:206944
 12:45 pm on Nov 27, 2003 (gmt 0)

Sandalwood - Yes Google do have dictionaries for different languages. I don't know how many yet.

And they do differentiate nouns, verbs, etc. That's why they need a dictionary - to hold the valid word variations.

sandalwood




msg:206945
 1:01 pm on Nov 27, 2003 (gmt 0)

Hmmmm.....

Have a look and after that tell me your opinion:

Search: "buy watch"
Have a look at #2 Result.

B0ok: Night W@tch
If this is G dictionary-grammar then G has discovered a new English language version....

Just Guessing




msg:206946
 1:13 pm on Nov 27, 2003 (gmt 0)

That one is nothing to do with an Enlish Language dictionary or word stemming, but it could be to do with a Commercial Search Terms database. It fits the conspiracy theory.

Marcia




msg:206947
 1:14 pm on Nov 27, 2003 (gmt 0)

So it's back to the same old thing some folks have harped on for ages - contextual relevancy, determining the meaning of a word by the context it's in.

Think bow: hair bow (noun), violin bow (noun), bow and arrow (noun), bow down (verb), bowed (as in bent - an adjective), bow as in bend from the waist similar to curtsey (verb).

sandalwood




msg:206948
 1:20 pm on Nov 27, 2003 (gmt 0)

Just_Guessing....please!

As a user, are you ok with the results of "buy watch"?
Come on, let's get serious.

The only explanation of the #2 result is the new G "algo":
The #2 is an Am@zon url!

kaled




msg:206949
 1:26 pm on Nov 27, 2003 (gmt 0)

Whilst it's impossible to go back in time and check, I would say that plurals have been used in searches for a very long time. However, only exact matches were displayed bold.

Kaled.

sandalwood




msg:206950
 1:30 pm on Nov 27, 2003 (gmt 0)

I would agree with the relation between plural and singular in search terms. But not in Grammar syntax...G can't make grammar syntax check and understand what is in user mind, and if some people say it can do it then G does it in a wrong way.

With all the above i have said before if you wanna buy a watch and search for it in G i think you will buy a Book which title includes the word "watch".
Ok, if you think about it it's gonna be cheaper for you..:)

[edited by: sandalwood at 1:33 pm (utc) on Nov. 27, 2003]

wruppert




msg:206951
 1:31 pm on Nov 27, 2003 (gmt 0)

A threshold of about $1 for top position fits with the results I have seen, but I don't want to believe the evidence!

I manage my wife's real estate site. "TownX Real Estate" has received the new treatment while "TownY Real Estate" has not. I am bidding $1.12 on both, I am third on both, and have the same two ads above me on both.

Just Guessing




msg:206952
 1:35 pm on Nov 27, 2003 (gmt 0)

sandalwood - No you misunderstand me - the new results are much much worse than they used to be for many search terms.

I'm just trying to figure out why - and as you say, it looks like there's lots of changes.

Word stemming is fraught with problems. Are these the same searches?

Blue Widgets
Widget Blues

sandalwood




msg:206953
 1:38 pm on Nov 27, 2003 (gmt 0)

Just_Guessing i can seperate the diffence...

But G can't pretend that can seperate the difference too...At least not in 2003!

And one question about that:
Does Google try to start a "New Generation" of Search Engines?

[edited by: sandalwood at 1:42 pm (utc) on Nov. 27, 2003]

Chris_D




msg:206954
 1:38 pm on Nov 27, 2003 (gmt 0)

Excellent find just_guessing!

It also explains why "possessive case" words are now highlighted in searches eg if you search for 'fred widget' - most results are actually returned for 'fred's widget' and fred's is highlighted.

killroy




msg:206955
 1:42 pm on Nov 27, 2003 (gmt 0)

The new filter (can be examined at ****) has removed some strange pages of mine...

The title is "Buy Widgets" and lists several sites and prices where you can buy widgets.

For a search for "buy widgets" where ti used t orank 3 with lots of other pages about buying widgets, all these pages have been removed. We're now left with:

"Buy widget-free Widgets" Wodgets that DO NOT include the ingredient widget. There is about 7 of th etop ten with these results, and a page with an articly about how "to buy spring brirds, I hit my shoulder on the widget in the corner" and that now ranks #3 for "Buy Widget".

Hardly a relevant result... hceck out that site which shows pre-florida results (I wish I knew how), and can aid in comparison and analysis.

SN

rogerd




msg:206956
 1:51 pm on Nov 27, 2003 (gmt 0)

I've noticed singular and plural searches for some words becoming similar lately. It never reached the point of Overture's identical results, at least for the keywords I checked. It also seemed a bit hard to reproduce when I was looking at this last week. I made a note to keep an eye on it. Perhaps there was some stemming testing taking place.

Just Guessing




msg:206957
 1:57 pm on Nov 27, 2003 (gmt 0)

I think 's was highlighted before, but plurals and other word stemming is new.

Word stemming seems to be coming and going at the moment. In fact I'm not seeing it at the moment, but I am still seeing 's. I assume it will come back because Google now say they do it in their Search Tips.

sandalwood




msg:206958
 2:06 pm on Nov 27, 2003 (gmt 0)

So WebMasters from now on feel free to use relative words on your kws.
How can we find that singular-plural-relative words list? ;)

CygnusX1




msg:206959
 4:38 pm on Nov 27, 2003 (gmt 0)

Hi Everyone,

I am grateful to all of your who have helped me in optimizing a website, and I donít want to step on anyoneís toes, but reading about some of the problems people have with search engines. I canít help but want to say something.

I have read all of the listings about Googleís new update and the fact that they are using a new stemming technology. I donít normally put my 2 cents in because I donít have the 6 or 7 years behind me to back up my statements.

I donít understand why so many SEOís are having trouble with the websites that they have optimized. All of the websites that I have optimized are all doing just fine. I canít help but wonder if too many SEOís are trying to bend the rules so much that once the rules get changed just a little that they no longer are within the rules. They find themselves across the line. Then they pay the price by dropping in listings or being penalized in some other way. When I optimize, I see where the boundary lines are and I get close to them, but I donít try and bend them. I canít help, but think that too many SEOís are going to far when they optimize a website.

Keep in mind that I have a great amount of admiration for those of you that helped me in learning about optimization.

I have had no trouble with any of my listings. I have had no trouble with the stemming technology. I have had no trouble with any of the Google updates in the past. Why is it that I donít have the problems that so many SEOís have? I truly believe that it is because a good amount of SEOís are actually spamming and not optimizing. I believe they forgot about ďOptimizing 101Ē and need to go back to the basics.

Thereís my 2 cents worth.

karembeu




msg:206960
 5:02 pm on Nov 27, 2003 (gmt 0)

for my 2 cents i've noticed that one of our main search terms 'kw1 kw2 kw3' has disappeared after Florida, but the plural version 'kw1 kw2 kw3s' is fine. Weird thing for me is that the singular term is in the page title and mentioned several times (legitimately) in the body. The plural is only mentioned once in the body, and not in the title.

Okay, here's the interesting bit. If I do my link:domain searches and check the link text of those pages, they all use the plural.

*I* think that the singular and plural SHOULD return the same, but the new Florida filter seems to think differently. I can get the singular term back up in the SERPS by bypassing the filter using any of the techniques mentioned elsewhere.

Just Guessing




msg:206961
 5:58 pm on Nov 27, 2003 (gmt 0)

Singular and plural should not return the same. Both should generate a match in the results, but an exact match will rank higher than the other singular or plural variation in the results (I believe).

karembeu - your experience of the Florida "filter" illustrates the filter theory very well. The theory is that the "filter" is applied only to certain commercial keywords (which may be why Cygnus doesn't have a problem), but this "over-optimisation filter" triggers on as little as an exact match on the keywords in the title plus as little as once or twice on the page. Your plural keyword escapes the filter but ranks highly because of incoming anchor text.

The big question is why Google thinks that keywords in the title and a couple of times on the page is over-optimisation, when it seems quite legitimate and normal to most people.

AthlonInside




msg:206962
 6:12 pm on Nov 27, 2003 (gmt 0)

The word stemming do exist and it is a fact, not an opinion, esp with singular/plural and the ...ing.

I am quite sure adwords broad match library are being used. Why need to build another set of word stemming database if you already got one? :)

Chndru




msg:206963
 6:23 pm on Nov 27, 2003 (gmt 0)

woohoo..this is cool stuff. me been wanting this stemming for a long time. I wish there is an option in the advanced search to turn it off, if i want. for e.g. i search for pant [google.com], but i want results only for this meaning [dictionary.reference.com].

this topic was one of my first posts [webmasterworld.com]

This 52 message thread spans 2 pages: 52 ( [1] 2 > >
Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Google / Google News Archive
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved