Feeding Arguments into an Optimization Function

Let's say you only want to optimize a few parameters in a function - not all of them. In this case, you should pass the rest of the parameters through arguments.

Example: The below function needs the data (time vector and observation vector) passed to it

def freqfitscore(inputs,timevec,obsvec):

Show full code for function

import math
import numpy as np

def freqfitscore(inputs,timevec,obsvec):
    R=inputs[0]
    fo=inputs[1]
    def freq(t,fo,R):
        #a:make sure exponential isn't too large
        ex=-R*t
        if ex>700.0:
            ex=700.0
        a=math.exp(ex)

        #b: second term
        b=(1.0-fo)
        #dm: write full denominator and make sure not zero
        dm=fo+b*a
        if dm==0:
            return 0.99
        else:
            return fo/dm

    
    predvec=[freq(t,fo,R) for t in timevec]
    predvec=np.array(predvec)
    diff=(obsvec-predvec)
    diffsq=diff*diff
    return sum(diffsq)

We import the function as "score" and pass the data to it during the optimization by using the arguments slot

a=fmin(score,guess,args=(timevec,obsvec),full_output=1)

Show full optimization code below

import numpy as np
from scipy.optimize import fmin
from freq_function import freqfitscore as score


#Data
#===============================================================================
##time points of each set of yes and no experiments (zeroed)
x=[0.0, 39.0, 82.0, 186.0, 255.0, 331.0]
#creating percentage of positive hits for each set
p=[0.0, 0.0, 0.05263157894736842, 0.42105263157894735, 0.8571428571428571, 0.9285714285714286]
#==================================================================================
#score a fit
timevec=np.array(x)
obsvec=np.array(p)
guess=(0.8,0.2)
a=fmin(score,guess,args=(timevec,obsvec),full_output=1)