寻找相关或共同变动的股票
我有一张每日收盘股票价格和商品价格（例如黄金、石油等）的表格。我想找到哪些股票与另一种股票或商品走势密切。
我从哪里开始进行这种类型的分析 - 我了解 java、SQL、python、perl 和一点点 R。
如有必要，愿意购买和学习 Matlab 等新工具。
任何指导都将受到高度赞赏。
这不是一个家庭作业问题。
谢谢..
I have a table of daily closing stock prices and commodity prices such as Gold, Oil, etc. I want to find what stocks move closely with another stock or a commodity.
Where do I start to do this type of analysis - I know java, SQL, python, perl, and a little bit of R.
Willing to buy and learn new tools like Matlab if necessary.
Any guidance will be highly appreciated.
This is not a homework question.
Thanks..
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（3）
您正在寻找的技术称为协整。在计算两个时间序列的协整时，语言根本不重要，因此请使用您喜欢的任何语言。
我不同意其他回答，认为计算不是问题。能够计算不同时间序列之间潜在的数十亿个协整系数是一个巨大的问题。使用高度优化的库至关重要。不过，这篇关于 R 中协整测试的文章应该可以帮助您入门。
另请访问 quant.stackexchange.com，了解有关量化金融的更多信息。
The technique you are looking for is called cointegration. Language is not important at all when computing cointegration of two time series so use whatever you are comfortable with.
I disagree with other responses that computation is not a problem. It is a huge problem to be able to compute potentially billions of cointegration coefficients between different time series. Using a highly optimized library is critical. However this article on cointegration testing in R should get you started.
Also checkout quant.stackexchange.com for more info on quantitative finance.
试试这个：
http://www.sectorspdr.com/correlation/
http://www.etfscreen.com/corr.php
http://correlate.googlelabs.com/faq
https://quant.stackexchange.com/questions/1027/correlation-和-协整-相似性-差异-关系/1038#1038
Try this:
http://www.sectorspdr.com/correlation/
http://www.etfscreen.com/corr.php
http://correlate.googlelabs.com/faq
https://quant.stackexchange.com/questions/1027/correlation-and-cointegration-similarities-differences-relationships/1038#1038
如果我是你，我会从 在 Google 学术搜索中搜索“comovement”一词。并非所有出现的内容都直接相关，但有相当多的内容是相关的。
通过浏览论文并进行更多谷歌搜索，您应该更清楚地了解要学习哪些类型的统计方法。
我同意 Ben Bolker 的观点，即计算工具并不是目前的主要问题。
If I were you, I'd start by searching Google Scholar for the word "comovement". Not everything that turns up is directly relevant, but there's quite a lot of stuff that is relevant.
By looking through the papers and googling some more, you should get a clearer picture of what types of statistical methods to learn.
I agree with Ben Bolker that computational tools are not the main issue at this point.