NBK wrote:Oh.. I guess 2m is not few enough to suit you. .... I shall endeavor to be more precise in my pointless posts in the future.
I was not trying to nitpick, but simply attempting to point out that I did not find the same abundance as suggested, and that among those results returned there on my computer there was no such proportion of significant examples.
NBK wrote:
You might notice them as the white or blue clothing. maybe there's a diff in which Google site is used.
Nope, nothing like that, no jûdôgi, no "white or blue clothing", nothing like that.
I think you are misunderstanding me. I think the returned results are in this case not so much the consequence of how we use of Google, but they are likely determined by your location. I suspect that given your location, it will also return more sites based on words that are synonyms. For that reason, the results we get using the same search using the same search engine may be quite different.
By, the way, you can also see a similar event when using YouTube. The movies depicted on the starting page there differ depending on the location recognized by the computer. Clearly, Google decides what to return based on IP address.
You will see this if you do a research using a search term of something that is present in every country and by preference using a word that is identical in several languages. An example of that is "restaurant". The returned results are then completely different between computers depending on your IP address. Your first page will guarantee included at least some sites that are located in the country in which your IP address is located.
I also notice that when I use kanji, that the returns I get are primarily Chinese; I believe this is not so if one's IP address is, for example, in another country that abundantly uses kanji, such as for example, Japan. This also has further consequences. Depending on a the location, Google may "interpret" the search term and include results for a "wrong" term by "guessing" what the correct term might be and by including synonyms for that term as may exist in that location. So, the fact that it returns X number of sites in response to a certain kanji does not imply that all those sites 'use' the kanji, but will likely include a considerable number of returns through synonym guessing, the number of which will differ by the location Google establishes you're in and the language spoken there, or at least, spoken there according to Google's interpretation.
In practical terms, I think that if me and you would pdf the first page of returns using the same search term and exchange or compare those pdf's that there would be considerable differences.
While I am not an expert in Internet, I did work for some time for a company that was a market leader in natural language software, and I worked precisely on data mining and how these things work with regard to medical databases including voice commands. When a physician, for example, dictated a recipe including the wrong doses, the software had to be able to question that and refuse or correct it. Moreover, the database was multilingual and I was, inter alia responsible for creation of the Japanese components. We also created databases with medical literature regarding the effectiveness of search engines. The rationale was that there was an abundance of information, and search had to be able to yield relevant results, so really what is about the definition of 'data mining'. I know that our technology was sold to some major US companies, and I do recognize some similarities in how Google functions hence my attention for these aspects without any intent to discredit those who make use of it.
So, no mischief behind my post, just assessing the structural reasons that may be responsible for leading to a certain result, which one might want to take into account when arriving at a conclusion. That's all.
Home
in BLUE!
