MLE¶
-
class
tweezepy.MLE.MCMC(walkers=32, steps=1600, progress=True, **kwargs)¶ Monte Carlo sampler class.
- Parameters
walkers (int, optional) – Number of walkers, by default 32
steps (int, optional) – Number of steps, by default 1600
progress (bool, optional) – Print progress bar, by default True
-
calc_mc_errors(percentiles=[15.87, 50, 84.13], discard=None, thin=None)¶ Computes percentiles from Monte Carlo samples.
- Parameters
percentiles (list, optional) – Percentiles for each parameter, by default [15.87,50,84.13]
discard (int, optional) – Number of “burn-in” steps to discard, by default 100
thin (int, optional) – N, by default 10
- Returns
errors – Errors from Monte Carlo sampling.
- Return type
array
-
corner_plot(quantiles=[0.16, 0.84], labels=None, **kwargs)¶ Utility function for generating corner plots.
- Parameters
quantiles (list, optional) – Quantiles to annotate, by default (0.16,0.84)
labels (list, optional) – Parameter labels, by default None
- Returns
fig (Figure) – Figure object.
axes (Axes) – Axes object.
-
sample_plot(fig=None, labels=[], fig_kwgs={}, ax_kwgs={})¶ Plot the accepted Monte Carlo samples.
- Parameters
fig (object, optional) – Figure object, by default None
labels (list, optional) – Plot labels, by default []
fig_kwgs (dict, optional) – Figure keywords, by default {}
ax_kwgs (dict, optional) – Axes keywords, by default {}
- Returns
fig (Figure) – Figure object.
axes (Axes) – Axes object.
-
class
tweezepy.MLE.MLEfit(pedantic=True, scale_covar=False, minimizer_kwargs={})¶ Perform maximum likelihood estimation and uncertainty calculations.
- Parameters
pedantic (bool, optional) – Ignore unhelpful warnings, by default True
scale_covar (bool, optional) – Whether to scale standard errors by reduced chi-squared, by default False
fit_kwargs (dict, optional) – Disctionary of keyword arguments passed to scipy.optimize.minimize, by default {}.
-
names¶ Fit function parameter names.
- Type
list
-
params¶ Parameter values.
- Type
array
-
std_errors¶ Parameter uncertainties.
- Type
array
-
chi2¶ Chi-squared value.
- Type
float
-
redchi2¶ Reduced chi-squared value.
- Type
float
-
AIC¶ Akaike information criterion.
- Type
float
-
AICc¶ Corrected Akaike information criterion.
- Type
float
-
mcmc(walkers=32, steps=2000, discard=100, thin=10)¶ Runs Monte Carlo sampler and computes standard errors as 0.5*(std_u - std_l)
- Parameters
walkers (int, optional) – Number of walkers, by default 32
steps (int, optional) – Number of steps to take, by default 2000
discard (int, optional) – Number of initial steps to discard, by default 100
thin (int, optional) – Distance between independent steps, by default 10
-
property
results¶ Dictionary of MLE fit results.
- Returns
Dictionary of MLE fit results.
- Return type
dict
-
class
tweezepy.MLE.Gamma_Distribution(shape, yhat)¶ Gamma probability distribution class.
- Parameters
shape (np.array) – Shape parameter.
yhat (np.array) – Experimental values.
-
shape¶ Shape parameters.
- Type
array-like
-
yhat¶ Experimnetal values
- Type
array-like
-
cdf(ytrue)¶ Cumulative distribution function.
- Parameters
ytrue (array-like) – True/theoretical values.
- Returns
cdf – Cumulative distribution function.
- Return type
callable
-
interval(ytrue, alpha=0.95)¶ - Parameters
ytrue (array-like) – True/theoretical values.
alpha (float, optional) – Probability that a random variable will be drawn from the returned range. Each value should be in the range [0, 1]., by default 0.95
- Returns
Interval – Endpoints of the range that contains fraction alpha [0, 1] of the distribution.
- Return type
callable
-
logcdf(ytrue)¶ Log of cumulative distribution function.
- Parameters
ytrue (array-like) – True/theoretical values.
- Returns
logcdf – Log of cumulative distribution function.
- Return type
callable
-
logpdf(ytrue)¶ Log of the probability distribution function.
- Parameters
ytrue (array-like) – True/theoretical values.
- Returns
logpdf – Log of the probability distribution function.
- Return type
callable
-
pdf(ytrue)¶ Calculate probability distribution function.
- Parameters
ytrue (array-like) – True/theoretical values.
- Returns
pdf – Probability distribution function.
- Return type
callable
-
scale(ytrue)¶ Calculate scale parameters.
- Parameters
ytrue (array-like) – True/theoretical values.
- Returns
scale – Scale parameters.
- Return type
array-like
-
std(ytrue)¶ Calculate standard deviation.
- Parameters
ytrue (array-like) – True/theoretical values.
- Returns
std – Standard deviations.
- Return type
array-like