这可能是个愚蠢的问题,但我在lmfit's documentation中找不到答案。我的问题很简单:如何检索R的平方?(我知道我可以用1 - SS_res / SS_tot手动计算)
更新:
我试着自己计算R平方,并将其与statsmodels中的R平方进行了比较。
两种估计中的参数是相同的,但R平方不是。在
代码:from lmfit import minimize, Parameters
import numpy as np
import as sm
import random
x = (0, 15, 10)
x_ols = sm.add_constant(x)
y = [(0,15) for r in xrange(10)]
model = (y,x_ols)
results = ()
print "OLS: ", format([0], '.5f'), format([1], '.5f'), "R^2: ",
# define objective function: returns the array to be minimized
def fcn2min(params, x, data):
a = params['a'].value
b = params['b'].value
model = a + b * x
return model - data
for i in range(0,1):
# create a set of Parameters
params = Parameters()
('a', value= i)
('b', value= 20)
# do fit, here with leastsq model
result = minimize(fcn2min, params, args=(x, y))
yhat = params['a'].value + params['b'].value * x
ybar = (y)/len(y)
ssreg = ((yhat-ybar)**2) # or sum([ (yihat - ybar)**2 for yihat in yhat])
sstot = ((y - ybar)**2) # or sum([ (yi - ybar)**2 for yi in y])
r2 = ssreg / sstot
print "lmfit: ", format(params['a'].value, '.5f'), format(params['b'].value, '.5f'), "R^2: ", r2