sorry for the delay i just noticed the reply , i'm more then willing to
share the code i'd love to participate in the project .
i've tried using
for (int i=19;i<list.size();i++){
//DescriptiveStatistics commonsStat = DescriptiveStatistics.newInstance();
SummaryStatistics commonsStat = SummaryStatistics.newInstance();
for (int j=i-19;j<=i;j++){
commonsStat.addValue(list.get(j).indicatorPrice());
}
Indicator indicator = new Indicator();
indicator.setTime(list.get(i).indicatorTime());
indicator.setValue(commonsStat.getStandardDeviation());
SD.add(indicator);
}
----- Original Message -----
From: "Phil Steitz" <[EMAIL PROTECTED]>
To: "Jakarta Commons Users List" <[email protected]>
Sent: Saturday, September 29, 2007 9:25 PM
Subject: Re: Commons.Math.Distrubition
On 9/29/07, Amnon Lahav <[EMAIL PROTECTED]> wrote:
Hi ,
how is the Standard deviation calculated in the commons math ? because
i'm
getting diffrent values then my code which i used till far , iswtiched
out
of runtime speed considerations and i know my code is OK since it mathces
the stock exchange's values ...
isn't is for 20days periods ?
thanks .
The exact computation depends on whether you supply the whole array of
values or feed them in one at a time; though for small arrays the
results should not differ. See
http://commons.apache.org/math/api-1.1/org/apache/commons/math/stat/descriptive/moment/Variance.html
for a description of the computational formulas used. Math.sqrt is
used to get the standard deviation from the variance.
What class / method are you using? Can you share the data and
expected results? If so and you can't explain the difference, please
open a Jira ticket here.
http://commons.apache.org/math/issue-tracking.html
Phil
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
