Mailman 3 March 2013 - SciPy-User

TODO: more one-sided options in statistical tests
by josef.pktd＠gmail.com March 2, 2013

March 2, 2013

options that would be nice to have, if someone has a few spare cycles. most statistical tests in scipy stats don't have an option for a one-sided alternative. I could use a one-sided ``binom_test`` right now. http://ugrad.stat.ubc.ca/R/library/stats/html/binom.test.html My guess is, it's at most of few lines, but it takes time to find them and to write the tests. For example, a ttest_ind that allows for one-sided alternatives and for a Null hypothesis that has a mean difference different from zero https://github.com/statsmodels/statsmodels/pull/535/files#L13R461 (I needed it to write an equivalence test) Josef

1 0

Spline interpolation array instead of loop?
by Michael Aye March 2, 2013

March 2, 2013

Hi! So, I have to do the same spine interpolation (same=same x-axis) for 189 data channels, so I was wonderding if there is a way to store the 189 measurements into a 2D array and to set that at once to an interpolator, is that possible somehow? I noticed that InterpolatedUnivariateSpline does not allow me to do that, it seems. So, in other words, what I would like to be able to do is: len(x) = 20 len(y_i) = 20 i = 1 ..189 y.shape == (20,189) s = UnivariateSpline(x, y) res = s(new_x) res.shape == (20, 189) Does something like that exist? Have a nice weekend! Michael

2 3

List of functions with different arguments
by Per Nielsen March 1, 2013

March 1, 2013

Hi all I apologize if this is too off topic, but I figured it was common in scientific computing. I am trying to make a list of copies of the same function, but with different extra input arguments. However, I am having problems getting the wanted output. I thought it was a problem with the way I refered to the function, but I tried a deepcopy and also the functions seem to have different ids. Here is my output from the attached script: simple [3, 3, 3] [4549677376, 4549677496, 4549677616] deepcopy [3, 3, 3] [4549677736, 4549677856, 4549677976] wanted/expected output: [1, 2, 3] Any help would be greatly appreciated :) Cheers, Per

3 2

_dop.error: failed in processing argument list for call-back fcn.
by SHIHUI GUO March 1, 2013

March 1, 2013

HI all, I want to use scipy to implement an oscillator, ie. solving a secod-order ordinary differential equation, my code is: =================================== from scipy.integrate import ode y0,t0 = [0, 1], 0 def fun(t, y, params): rou = 1 omega = 10 sigma = 1 # convergence rate, ie, lambda conrate = 10 temp = -conrate*((y[0]^2+y[1]^2)/rou^2-sigma) dy = temp*y[0] - omega*y[1] ddy = omega*y[0] + temp*y[1] return [dy, ddy] test = ode(fun).set_integrator('dopri5') test.set_initial_value(y0, t0) t1 = 10 dt = 0.1 while test.successful() and test.t<t1: test.integrate(test.t+dt) print test.t, test.yy =================================== =================================== The error says: _dop.error: failed in processing argument list for call-back fcn. File "/home/shepherd/python/research/testode.py", line 23, in <module> test.integrate(test.t+dt) File "/home/shepherd/epd/epd_free-7.3-2-rh5-x86/lib/python2.7/site-packages/scipy/integrate/_ode.py", line 333, in integrate self.f_params, self.jac_params) File "/home/shepherd/epd/epd_free-7.3-2-rh5-x86/lib/python2.7/site-packages/scipy/integrate/_ode.py", line 827, in run tuple(self.call_args) + (f_params,))) =================================== Previously I use the ubuntu default python, and scipy is 0.9.0. Some thread says it is a bug and has been fixed in 0.10.0, so I switched to enthought, now the scipy is newest version, but the error remains. Thanks for any help. Shihui -- * --------------------------------------------------------------------------------------- * SHIHUI GUO National Center for Computer Animation Bournemouth University United Kingdom BU is a Disability Two Ticks Employer and has signed up to the Mindful Employer charter. Information about the accessibility of University buildings can be found on the BU DisabledGo webpages [ http://www.disabledgo.com/en/org/bournemouth-university ] This email is intended only for the person to whom it is addressed and may contain confidential information. If you have received this email in error, please notify the sender and delete this email, which must not be copied, distributed or disclosed to any other person. Any views or opinions presented are solely those of the author and do not necessarily represent those of Bournemouth University or its subsidiary companies. Nor can any contract be formed on behalf of the University or its subsidiary companies via email.

2 4

optimize.brentq and step function
by josef.pktd＠gmail.com March 1, 2013

March 1, 2013

brentq documentation says "f must be a continuous function" I forgot that I have a step function and tried brentq and it worked. Is this an accident or a feature? the underlying function has a ceil and floor in it to convert the argument to integer. example: trying to invert stats.binom_test to get a confidence interval >>> n_rep 10000 >>> for count in np.arange(455, 460, 0.5): count, stats.binom_test(count, n_rep, p=0.05) ... (455.0, 0.038914401175656158) (455.5, 0.038914401175656158) (456.0, 0.043471737392089219) (456.5, 0.043471737392089219) (457.0, 0.048471945345671313) (457.5, 0.048471945345671313) (458.0, 0.053946450095604122) (458.5, 0.053946450095604122) (459.0, 0.059927554205846612) (459.5, 0.059927554205846612) >>> def func(qi): ... return stats.binom_test(qi * n_rep, n_rep, p=0.05) - 0.05 ... >>> qi = optimize.brentq(func, 0.01, 0.05) >>> (qi * n_rep) 457.99999999456901 >>> stats.binom_test(np.floor(qi * n_rep), n_rep, p=0.05) 0.048471945345671313 >>> stats.binom_test((qi * n_rep), n_rep, p=0.05) 0.048471945345671313 >>> stats.binom_test(np.floor(qi * n_rep)+1, n_rep, p=0.05) 0.053946450095604122 It looks like I get an integer next to the step root. Josef

2 3