[SciPy-User] leastsq returns bizarre, not fitted, output for float values

[Matthieu Rigal]

Hi folks,

Thanks for your help last time, even if I had not reply to my second 
message...

I am using leastsq for several things, but it is returning strange values 
for one of the case I'm using it for. I simplified it to the following 
code I'll paste below. 
The effect is that it is fitting nothing, just giving back the parameters 
given for initialization. Thus some fitting is possible as you will see in 
the plotted graph.
I'm using SciPy 0.7.

It might be a bug, a misusage from my side... or some data type 
incompatibility I was not able to find on the net or in the source...

As you will see, if you transform the x-data to a numpy.int array (by 
uncommenting a line below), the fitting is working... is it to be 
expected ? It should then be somewhere in the doc, isn't it ?


import numpy
from scipy.optimize import leastsq
import matplotlib.pyplot as plt

def LinearFit(p, y, x):
    a, b = p
    return y - (a*x +b)

aX = numpy.asarray([ 22.08742332,  23.43987274,  21.59165192,  24.80192566,
        26.11182976,  29.18944931,  27.89473152,  30.00043106,
        36.24227142,  30.45967293,  30.04778099,  28.11702538,
        29.31716728,  27.89473152,  20.59804916,  34.19070053,
        48.33156204,  50.82163239,  45.22343063,  42.80136108,
        30.71160889,  29.31716728,  25.14836884,  23.50605965,
        26.89011765,  40.35306168,  55.074543  ,  58.57307816,
        60.77198792,  56.14603043,  39.29994583,  38.14756012,
        35.76476288,  27.31066895,  23.45325851,  30.46047974,
        37.53346634,  41.04254532,  54.47524643,  61.14104462,
        61.03421402,  56.14603043,  44.67305756,  35.13313675], 
dtype=numpy.float32)

aY = numpy.asarray([ 25.45091248,  25.50468063,  27.15722656,  25.10549927,
        28.44662094,  30.3882637 ,  31.90523148,  34.12581253,
        36.62049484,  33.90032196,  34.04083252,  29.66094398,
        30.68564224,  29.31051826,  25.17509079,  37.28609848,
        42.86494827,  48.25041199,  46.88908005,  34.44023132,
        31.26217461,  31.8005867 ,  28.34657669,  26.77126312,
        31.06710815,  41.03251266,  49.48557281,  52.79579926,
        50.865448  ,  48.03937531,  39.30026245,  38.50889969,
        37.07154083,  31.61130905,  27.42698288,  30.84166718,
        30.84166718,  40.47367096,  50.37258148,  53.13900757,
        53.75816727,  52.74428177,  43.87319183,  33.70808029], 
dtype=numpy.float32)
        
#aX = numpy.asarray(numpy.rint(aX), dtype=numpy.int)

p0 = [1.0] + [aY.min()-aX.min()]
aParams, err, i, j, k = leastsq(LinearFit, p0,  args=(aY, aX), 
maxfev=10000,full_output=True)
aY0 = aParams[0] * aX + aParams[1]
print err, i, j, k 
print aParams

plt.plot(aX, aY, '+', aX, aY0, '+')
plt.legend(['input','model'])
plt.show()


Thanks in advance for the help,
Best Regards,
M

-- 
Matthieu Rigal

RapidEye AG
Molkenmarkt 30
14776 Brandenburg an der Havel
Germany
 
Follow us on Twitter! www.twitter.com/rapideye_ag
 
Head Office/Sitz der Gesellschaft: Brandenburg an der Havel
Management Board/Vorstand: Wolfgang G. Biedermann
Chairman of Supervisory Board/Vorsitzender des Aufsichtsrates: 
Juergen Breitkopf 
Commercial Register/Handelsregister Potsdam HRB 17 796
Tax Number/Steuernummer: 048/100/00053
VAT-Ident-Number/Ust.-ID: DE 199331235
DIN EN ISO 9001 certified
 
*************************************************************************
Diese E-Mail enthaelt vertrauliche und/oder rechtlich geschuetzte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese
E-Mail irrtuemlich erhalten haben, informieren Sie bitte sofort den
Absender und vernichten Sie diese E-Mail. Das unerlaubte Kopieren sowie
die unbefugte Weitergabe dieser E-Mail ist nicht gestattet.
 
The information in this e-mail is intended for the named recipients
only. It may contain privileged and confidential information. If you
have received this communication in error, any use, copying or
dissemination of its contents is strictly prohibited. Please erase all
copies of the message along with any included attachments and notify
RapidEye AG or the sender immediately by telephone at the number
indicated on this page.