Google and Ye Shall Find???

Here are the results of the class runoff between Google and Yahoo (Assignment 6):

Measure: precision within first 20 results

number of queries for which Google had better precision: 8
number of queries for which Yahoo had better precision: 5
ties: 2

Google’s average precision at 20: 65% (standard deviation 24.4)
Google’s median precision at 20: 65%
Google’s highest precision at 20: 100%
Google’s lowest precision at 20: 15%

Yahoo’s average precision at 20: 61.667% (standard deviation 22.7)
Yahoo’s median precision at 20: 65%
Yahoo’s highest precision at 20: 100%
Yahoo’s lowest precision at 20: 15% (not on same query as Google’s lowest)

average difference (absolute value) between Google’s precision and Yahoo’s precision for a query: 14.667

Measure: rank of third relevant result

number of queries for which Google was better: 6
number of queries for which Yahoo was better:4
ties: 5

Google’s average rank of 3rd relevant: 4.6 (standard deviation 3.2)
Google’s median rank of 3rd relevant: 3
Google’s best rank of 3rd relevant: 3
Google’s worst rank of 3rd relevant: 14

Yahoo’s average rank of 3rd relevant: 5.2 (standard deviation 2.7)
Yahoo’s median rank of 3rd relevant: 4
Yahoo’s best rank of 3rd relevant: 3
Yahoo’s worst rank of 3rd relevant: 12

Given these numbers, it is fair to conclude that Google has performed better in this “runoff”. Yahoo’s worst rank of the 3rd relevant result (12) is better than Google’s (14); these were achieved on the same query. Otherwise, Google does as well or better than Yahoo under every measure. However, the numbers are not that far apart.

A few students tried to characterize Google’s results versus Yahoo’s in a general way, e.g. broad versus specific. However, there was no consensus. Read the individual entries to find out more.

With the Golden Globe nominations released Thursday, the season for movie awards has begun. A series of critics and guild awards will lead up to the Oscars in February. Thinking about this made me a bit nostalgic and reflective, so I decided to search for “most memorable Oscar speeches.” I know there have been some good ones, so the search engines should have been able to provide some relevant results. Both did a pretty good job, with Google scoring 75% and Yahoo scoring 65%, although I was a bit lenient with what I accepted as relevant. Both search engines provided some lists of the greatest Oscar moments, which were mostly speeches, so I accepted those, as they were what I was looking for the most. However, if a result focused on a specific memorable moment I accepted that too, because it did help me with what I was looking for. The most common instances of this happening were news stories from recent years focusing on the most memorable speech of the night the day after the ceremony. The irrelevant results that got mixed in were mostly musings from blogs or news sites about what made a good speech or how to write a good speech. Those were not helpful in my search for actual good speeches from the past. It took Google 5 results to get 3 relevant ones, and it took Yahoo 4 results.

For my search test I chose the query “black hatting search engines,” as it covers the topic of my final paper. In judging relevance I took a fairly open stance; anything that provided any sort of information on black hatting search engines, including opinionated blogs on the topics, all qualified as “relevant,” as they have information that I deem valuable to my research. I found that Google had a 100% hit rate, and Yahoo 95%. Yahoo almost had a 100% rate, but one of the links took me to a news website that had long since gotten rid of the article that contained the information on my query – leaving me with a blank page with no information. The rank of the 3rd relevant result for both search engines was 3. Both engines did very well with respect to this query.

My query for this exercise was ‘best presidential candidates.’ I found some pretty funny results that came from this relevant query considering the election that is coming up. On three of the four engines, one of the top results for this was a statement from the Israeli Panel, that they think Guiliani would be the best president for the US when the elections roll around. However Google was the only very helpful engine in terms of consistent results between one and ten that were relevant to my search. All of the results, although different, addressed this upcoming election, and also addressed the most important issue for my query, whos the best? They answered this question by giving me news links, polls, and leading candidates on specific issues. The other engines had results in their top ten that focused on one of the three words in my query, rather than addressing them as a phrase. I got many results from Ask that were not relevant, either talking about the presidency, or simply giving articles where the candidates were mentioned. One aspect of Ask’s results page that I don’t like is the fact that they put sponsored links above and below the standard results, making it more confusing, and in general focusing more on advertising than just getting people the results they need.

My query is ‘american gangster quotes’ and my criteria for success is whether or not the link given gives me some kind of quotes from the movie American Gangster. When I searched this through Google and Yahoo, very few of the results actually gave me quotes from the movie. Google gave me seven of the twenty results that had quotes in it, while yahoo only had three. Therefore the percentages out of the top twenty for Google and Yahoo are 35% and 15% respectively. Yahoo took until its fifth result to get its third relevant result where Google got its third relevant result from its fourth overall result.

I searched for the query “top 100 universities” (no quotations in the query), and deemed results relevant if they provided a ranking of colleges/universities, regardless of methodology behind the ranking or location of the institutions. A site that described one or more universities on the list but did not provide a ranking list was irrelevant. Google’s first 3 sites were relevant, while Yahoo did not return a 3rd relevant site until the 5th result. In addition to a better top 3, Google had 75% relevance in the top 20 results, while Yahoo had a close but slightly worse relevance of 70%.

For this search, I used the keywords “Russia media rights” (no quotes) because that’s what I’m currently writing about for my politics class. I started off with “media freedom Russia” and ended up with 100% accuracy on both search engines. With my second search, I had the same criteria for relevancy, but found completely different results.

Google precision of first 20 results: 90%
Google rank of third relevant result: 3
Yahoo! precision of first 20 results: 50%
Yahoo! rank of third relevant result: 10

On Yahoo! the first two results were actually irrelevant to my search, but that’s because they focused on “rights” and “Russia” and left out sites specifically about media rights. If I were unsure of what I was looking for, I might actually turn to Yahoo for a mix of results; for a very specific topic like a paper, however, Google came out on top.

            Comparing the query results for Google and Yahoo’s search engines showed expected conclusions, Google is superior. I used the query term “Spanish Geography” and %50 of Google’s results I found to be relevant while %40 of Yahoo’s were. Additionally the rank of the 3  relevant term was higher for Google (3) and lower for Yahoo (7).

            I felt that a relevant site was one which would tell any person who knew nothing about Spanish geography, a fact about the topic. I also felt the query term was fair being not too hard or too difficult. I actually felt it was a little more on the easy side and expected a higher percentage of relevant result for both search engines. The main problem both encountered was geography quizzes which had no information about the geography of Spain on the site. These appeared numerous times for both Google and Yahoo. Other problems included geography sites with information not pertaining to
Spain.

          Taking this test further I looked at the 91-100^th results for both search engines out of curiosity. It was interesting to find that 7/10 of Google’s results were relevant and only 1/10 of Yahoo’s results were relevant. Google was still finding new sites which had information pertaining to the query term. Yahoo continued to find geography quizzes and even had some results which had to do with Spanish food. While few people if any look at this number of results for a query page, it just goes to show how Google’s algorithm is currently superior to Yahoos.  �

My search query was: “Writing and physical health benefits”(no quotes on actual query).  This search was inspired by a paper that I had to write for one of my classes about writing.  In order for a site to be relevant I expected it to contain information specifically about how writing affects physical health. Not just simply be about writing about physical health or just contain both writing and physical health in a totally unrelated way.

Both Google and Yahoo returned 18 relevant sites out of the first twenty for 90% precision.  The first three sites on both engines were relevant, so the rank of the third relevant site was obviously a tie as well.  The only possible difference between the sites is that the 4th site on Yahoo was unrelevant for my search while the 11th site on Google was the first unrelevant one.  This arguably gives google an edge, but based on our criteria the two engines were an exact tie.�

In my hunt for information on a recent research paper for my writing seminar, I wanted to know about the different views and background information of the “Robber Barons” of the Gilded Age. Optimists viewed the Robber Barons as “Captains of Industry” due to the contributions they made on the early economy of America during the end of the 19th century and the construction of the Transcontinental Railroad. I looked for any information on these two views as well as examples of these views through the entrepreneurs dubbed Robber Barons. Some relevant sites had some background information and included several relevant links where more detailed information could be found relating to the issues they covered on their own site.
When I searched the query “Robber Barons Captains of Industry”, Google and Yahoo yielded very different results. The second and third ranked results Google showed were relevant information on the definitions of Robber Barons and their history. I found these sites relevant since they gave ample, detailed information on who they were and their impact on society. However, the following six results were much less helpful. They contained class assignment and outlines for college and high school courses, as well as two “cheat” sites that offered to write an essay on the topic for you at a cost. Scattered among the remaining results were a mix of pertinent articles and history lessons along with one book from Amazon, which could not be viewed, and what appeared to be a high school worksheet on the Robber Barons. The largest cluster of relevant information, strangely, appeared on the last six results. Unfortunately, since most people, I for one, usually stick to using the top 10 results on the first page, these useful links might often go unnoticed. This would have been useful if I had not already done most of my research on the topic prior to this assignment. Overall, 9 out of the top 20 results were deemed relevant, a surprising 45%.
Yahoo’s results represented some of the same relevant sites but in a drastically different ranking. Yahoo’s top four results contained the similar relevant as that of Google’s second and third. These results proved to contain useful information about the topic. The major difference of the two search engines lies in Yahoo’s next six results. Yahoo remained more consistent in displaying more information based web pages. This proved more useful since these first 10 results remained on the first results page. The second page was mixed with similar high school assignment and course outline pages that Google displayed with a higher ranking. However, once again there were three significantly relevant pages ranked 17th to 19th that I feel would have been much more useful for research had they been ranked on the first page. Yahoo’s relevance bolstered 13 relevant results from the top 20, yielding a 65% relevance of results. Surprising outdoing the search giant, Google.
What I drew from this assignment is that there is no flawless search engine. They all yield different results dependent on the query term. Also, it would be a wise decision to scour more than the first page of results for pertinent information because, after all, it may prove worthwhile to search a little more for much greater information that can be used for research.