使用Scipy.optimization的蛮干功能时，为什么会出现“ ValueError：设置具有序列的数组元素”的问题？

Question

Working on a model to predict electrophysiological data given a set of parameters. This script is trying to find values for those parameters that give predictions closest to experimental data. I'm running Python 2.7, Scipy 0.17.0, and Numpy 1.10.4. The script is attached below. The line that is getting the error is epsc_sims[n,1] = r_prob*poolsize .

Here is the script:

import scipy.optimize as optimize
import numpy as np
import math

def min_params(*params):
    std_err = 0
    epsc_exp = np.loadtxt('sample.txt')
    max_pool = params[0]
    r_prob = params[1]
    tau_recov = params[2]
    poolsize = epsc_exp[0,1]/r_prob
    epsc_sims = np.copy(epsc_exp)
    count = epsc_exp.size

    for n in xrange(1 , count/2):
        poolsize = poolsize - epsc_sims[n-1, 1]
        poolsize = max_pool + (poolsize - max_pool) * math.exp((epsc_sims[n-1, 0] - epsc_sims[n,0]) / tau_recov)
        epsc_sims[n,1] = r_prob*poolsize
        std_err += (epsc_exp[n,1] - epsc_sims[n,1])**2

    std_err /= count
    return std_err

params = (1e-8, 0.2, 0.5)
rranges = (slice(5e-9,5e-8,1e-9), slice(0.1, 0.3, 0.01), slice(0.3, 0.4, 0.01))
y = optimize.brute(min_params, rranges, args = params)
print y

And here is the Traceback (most recent call last):

Traceback (most recent call last):

  File "<ipython-input-25-21d343f36a44>", line 1, in <module>
    runfile('C:/Users/brennan/Google Drive/Python Scripts/Inhibitory Model/brute.py', wdir='C:/Users/brennan/Google Drive/Python Scripts/Inhibitory Model')

  File "D:\Python\Anaconda2\lib\site-packages\spyderlib\widgets\externalshell\sitecustomize.py", line 699, in runfile
    execfile(filename, namespace)

  File "D:\Python\Anaconda2\lib\site-packages\spyderlib\widgets\externalshell\sitecustomize.py", line 74, in execfile
    exec(compile(scripttext, filename, 'exec'), glob, loc)

  File "C:/Users/brennan/Google Drive/Python Scripts/Inhibitory Model/brute.py", line 33, in <module>
    y = optimize.brute(min_params, rranges, args = params)

  File "D:\Python\Anaconda2\lib\site-packages\scipy\optimize\optimize.py", line 2604, in brute
    Jout = vecfunc(*grid)

  File "D:\Python\Anaconda2\lib\site-packages\numpy\lib\function_base.py", line 1811, in __call__
    return self._vectorize_call(func=func, args=vargs)

  File "D:\Python\Anaconda2\lib\site-packages\numpy\lib\function_base.py", line 1874, in _vectorize_call
    ufunc, otypes = self._get_ufunc_and_otypes(func=func, args=args)

  File "D:\Python\Anaconda2\lib\site-packages\numpy\lib\function_base.py", line 1836, in _get_ufunc_and_otypes
    outputs = func(*inputs)

  File "D:\Python\Anaconda2\lib\site-packages\scipy\optimize\optimize.py", line 2598, in _scalarfunc
    return func(params, *args)

  File "C:/Users/brennan/Google Drive/Python Scripts/Inhibitory Model/brute.py", line 25, in min_params
    epsc_sims[n,1] = r_prob*poolsize

ValueError: setting an array element with a sequence.

The text file I use for spikes = np.loadtxt('sample.txt') is formatted as follows with ~3,000 lines:

0.01108 1.223896e-08
0.03124 6.909375e-09
0.074   6.2475e-09
0.07718 3.895625e-09

This is my first post on here so please let me know if I need to change anything or provide more info!

Answer 1

The scipy.optimize routines call the function with a vector of parameters to be optimized. Your function is thus called as min_params(x, *params) , where the *params are your custom arguments to the function that you supplied to the function using the keyword argument args . The way you define your function x will end up as the first elements of params inside the function.

Assuming that max_pool , r_prob , tau_recov are what you want to optimize over here is how to fix things:

def min_params(params):
    ...
y = optimize.brute(min_params, rranges)