An Interesting CS Ranking Based On Google

原文
http://pages.cs.wisc.edu/~remzi/rank.html


Introduction: Rankings of Computer Science and Engineering departments continue to be a contentious topic. What factors should be counted, and how should they be weighted? Ask a different person, get a different answer. Can we do better?

The answer is a resounding yes. We propose a simpler, direct, and likely more meaningful method: Google-based Ranking. By building on top of the years of technology put into Google's novel PageRank™ algorithm, we can leverage the power of the Internet and deliver a complete ranking of the top departments in the field, all with the click of a button.

Methodology: We typed the following search terms into Google: Computer Science, Computer Sciences, and Computer Engineering. The second plural form of Computer Science was included because some departments (particularly the author's own institution) are known as a Department of Computer Sciences for strange and historical reasons. We then copied the top 30 results, excluding links to Wikipedia and other non-academic institutions and advertisements, thus leaving 26 or 27 schools. The results are listed below. We include, for the sake of comparison, the most recent U.S. News and World Report rankings.


We take the top 10 schools from the U.S. News Computer Science rankings and sort them by each of our Google Ranks.

U.S. News Top 10 Computer Science departments, sorted by Google: Computer Sciences ranking:

1. Wisconsin
2. Texas
3. Stanford
4. MIT
5. CMU
6. Berkeley
7. Princeton
8. Illinois
9. Cornell
10. Washington

U.S. News Top 10 Computer Science departments, sorted by Google: Computer Science ranking:

1. CMU
2. Berkeley
3. Stanford
4. Cornell
5. Illinois
6. Princeton
7. Wisconsin
8. Texas
9. Washington
10. MIT

Given this data, we now compute our own, novel AverageRank™ statistic, which simply averages the above two tables and ignores virtually all other data, in what some have called a "useless" and others a "flat out wrong" statistical method:

1. CMU
1. Stanford
3. Wisconsin
4. Berkeley
5. Texas
6. Princeton
6. Illinois
6. Cornell
9. MIT
10. Washington

Conclusions: From our study, assuming that Google "knows all" and is likely right, we draw three primary conclusions:

    * Wisconsin is likely underrated by U.S. News. This result is no surprise, particularly given the author of the article.
    * MIT and Washington are likely overrated by U.S. News. This point is probably more surprising. Perhaps scholarship such as this and this has led to the demise of these highly-regarded schools.
    * Other schools are about in the right place. This point is especially true if you (1) ignore some results and (2) play with the data in an odd and unsound manner, using (for example) the much-maligned AverageRank™ algorithm.

Future Work: In the future, we plan to conduct other searches, using advanced Google features such as quotes around the search terms and the OR keyword. We also hope to automate this process even further, thus eliminating the need to type up a report or even look at this web page again.

lol 原来这人自己是uwm，不过他在伯克利的Advisor居然是David A. Patterson。不过page rank不是和访问量还有网页上的东西有关吗？

我日，这人的老婆居然也是uwm的cs教授。cs教授和cs教授结婚是噩梦啊～～

我日，这人的老婆居然也是uwm的cs教授。cs教授和cs教授结婚是噩梦啊～～
edwardgtxy 发表于 2009-11-6 22:28

这种情况还不少.

你查看一下UMass Laser, 看看那两口子, 那才叫类似. 不过人家两个人都很好啊, 这种夫妻店多几个也不错.

原文
http://pages.cs.wisc.edu/~remzi/rank.html


Introduction: Rankings of Computer Science and Engineering departments continue to be a contentious topic. What factors should be counted, and how ...
Xeen 发表于 2009-11-6 21:53

多谢分享!

关于ranking,以前有人写过正式论文发表,仔细捣鼓以下, 发点论文也没问题