从这段代码中,我可以打印最终符合" out.best_fit",我现在要做的是将每个峰绘制为单独的高斯曲线,而不是全部它们合并在一条曲线上。
from pylab import *
from lmfit import minimize, Parameters, report_errors
from lmfit.models import GaussianModel, LinearModel, SkewedGaussianModel
from scipy.interpolate import interp1d
from numpy import *
fit_data = interp1d(x_data, y_data)
mod = LinearModel()
pars = mod.make_params(slope=0.0, intercept=0.0)
pars['slope'].set(vary=False)
pars['intercept'].set(vary=False)
x_peak = [278.35, 334.6, 375]
y_peak = [fit_data(x) for x in x_peak]
i = 0
for x,y in zip(x_peak, y_peak):
sigma = 1.0
A = y*sqrt(2.0*pi)*sigma
prefix = 'g' + str(i) + '_'
peak = GaussianModel(prefix=prefix)
pars.update(peak.make_params(center=x, sigma=1.0, amplitude=A))
pars[prefix+'center'].set(min=x-20.0, max=x+20.0)
pars[prefix+'amplitude'].set(min=0.0)
mod = mod + peak
i += 1
out = mod.fit(y_data, pars, x=x_data)
plt.figure(1)
plt.plot(x_data, y_data)
plt.figure(1)
plt.plot(x_data, out.best_fit, '--')
全球契合的情节:
答案 0 :(得分:1)
我认为你想在健身后做到这一点:
SELECT t.pat
FROM `tab` t
JOIN `tab` r
WHERE r.pat = 'AA' -- input
AND t.path LIKE CONCAT(r.path, '/', '%')
AND t.has_tree = 1;
也就是说,复合模型的has_tree
将返回一个字典,其中包含作为组件模型前缀的键,以及作为该组件的计算模型的值。