Re: [boost] [math] Empirical distribution function
2011/6/15 <boost-request@lists.boost.org>
Date: Wed, 15 Jun 2011 09:35:49 -0400 From: er <er.ci.2020@gmail.com>
distributions< http://www.boost.org/doc/libs/1_46_1/libs/math/doc/sf_and_dist/html/math_too... . But an accumulator is just fine for now. By the way, does it mean that if we have a set of 1 million of distinct real numbers, the corresponding accumulator will store a pretty big number of counters?
Only if the million are distinct values. Otherwise it keeps each distinct value + a count.
Ok, that's how I understood it, then (even though I'm not sure that that's the way accumulators are supposed to be used).
I'm glad you like the idea of a wrapper on top of an accu, but I only plan to do the accu. Please give me a bit of time to do a bit of maintenance work, though. I'll keep you updated.
It sounds perfect: thanks! D
Hope this can serve as a basis for a conversation: https://svn.boost.org/svn/boost/sandbox/acc_ecdf/ I hope Eric Niebler is fine with the copyrights. If not, I'll modify as he sees fit.
I've updated this a bit. Here's the link to the doc: https://svn.boost.org/svn/boost/sandbox/accumulator_ecdf/libs/accumulators/e... It computes fairly straightforward but commonly used statistics .e.g typedef double float_; accumulator_set< int, stats<ecdf::tag::cdf<float_> >
acc;
acc( 1 ); acc( 2 ); acc( 2 ); acc( 3 ); acc( 3 ); acc( 3 ); // percentage of sample observations less than equal to 1: float_ y = ecdf::cdf<float_>( acc , 1 ); Also note in directory /aux_ there is an auxiliary tool for assessing convergence of a random generator to a stated statistical distribution. Here's an example: https://svn.boost.org/svn/boost/sandbox/accumulator_ecdf/libs/accumulators/e... Please send feedback. Thanks.
participants (2)
-
Denis Arnaud -
er