Package qubx :: Module fit

Module fit

Curve Fitting.

QUB Express is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

QUB Express is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License, named LICENSE.txt, in the QUB Express program directory. If not, see <http://www.gnu.org/licenses/>.

Classes

[hide private]

BaseCurve
Defines the interface of a curve object; curves need not subclass, but they must behave like it.

BaseFitter
Defines the interface of a fitter object; fitters need not subclass, but they must behave like it.

Curve
Fittable curve from a parsed Python expression; either simple algebraic, or system of ODEs; see acceptODEs.

DblDecExpoCurve

DblGaussianCurve

DblGaussianUnitCurve

DecExpoCurve

GaussianCurve

GaussianUnitCurve

HillCurve

HillCurveNorm

LMFitter
Minimizes sum-squared-residual using lmmin from qubx.fast.

LinearCurve

LogExpoCurve

LogExpoCurve2

LogExpoCurve3

LogExpoCurve4

LogExpoCurve5

LorentzianCurve

OutOfBounds

QuadGaussianCurve

QuadGaussianUnitCurve

QuintGaussianCurve

QuintGaussianUnitCurve

SimplexFitter
Minimizes the sum-squared-residual using scipy.optimize.fmin.

Simplex_LM_Fitter
Minimizes sum-squared-residual using SimplexFitter then LMFitter; using half the max_iter for each algorithm.

Stop

TriGaussianCurve

TriGaussianUnitCurve

Functions

[hide private]

Correlation(curve, param_vals, xx, yy, vvv, ww)
Returns (cross_correlation, is_pseudo, std_err_est, ssr, r2) cross_correlation matrix[i][j] for curve param indices i,j is_pseudo whether Hessian was non-positive-definite, requiring Cholesky factorization std_err_est error estimates for each curve param ssr sum-sqr residual r2 R^2

source code

Covariance(curve, param_vals, xx, yy, vvv, ww)
Returns the covariance matrix[i][j] (inverse Hessian), for curve param indices i,j.

source code

Hessian(curve, param_vals, xx, yy, vvv, ww, dev=1e-07)
Returns the numpy.array(shape=(i,j)) of d(dResidual)/dP_i,dP_j, for i,j in len(non-const-params).

source code

RunsProb(yy, ff)
Counts the number of runs of y>f and y<f; returns probability it's from random noise.

source code

calc_r(xx, logbase=10.0)

source code

main()

source code

on_iter_pass(param_vals, iter)
Default fitter iteration callback; does nothing.

source code

on_status_print(msg)
Default fitter status callback; prints to stdout.

source code

pool_sub_eval(args)

source code

pseudovar(var)
Minimally after Gill-Murray: if covariance matrix is not positive definite, some diagonals are negative, with unknown meaning.

source code

sub_eval(xx, vvv, func, pval_by_name, avail)

source code

Variables

[hide private]

cpu_count = 4

erfc = <ufunc 'erfc'>

Curves = {'AllExponentials': <class 'qubx_single.dur_hist.AllE...

D1_DIFF_CENTRAL = False

D2_DIFF_CENTRAL = True

DIFF_DEV = 1e-07

Fitters = {'Levenburg-Marquardt': <class 'qubx.fit.LMFitter'>,...

GaussianCurves = [<class 'qubx.fit.GaussianCurve'>, <class 'qu...

GaussianUnitCurves = [<class 'qubx.fit.GaussianUnitCurve'>, <c...

USE_POOL_THRESHOLD = 1024

__package__ = 'qubx'

Function Details

[hide private]

Hessian(curve, param_vals, xx, yy, vvv, ww, dev=1e-07)

source code

Returns the numpy.array(shape=(i,j)) of d(dResidual)/dP_i,dP_j, for i,j in len(non-const-params).

Returns:: (hessian_matrix, sum_sqr_residual)

pseudovar(var)

source code

Minimally after Gill-Murray: if covariance matrix is not positive definite, some diagonals are negative, with unknown meaning. here we look for trouble and if necessary replace the covariance matrix with a generalized cholesky factorization.

Variables Details

[hide private]

Curves

Value:

{'AllExponentials': <class 'qubx_single.dur_hist.AllExponentials_Displ
ay'>,
 'Custom': <class 'qubx.fit.Curve'>,
 'Declining Exponential': <class 'qubx.fit.DecExpoCurve'>,
 'Double Declining Exponential': <class 'qubx.fit.DblDecExpoCurve'>,
 'Exponential (log x)': <class 'qubx.fit.LogExpoCurve'>,
 'Exponential (log x) (2)': <class 'qubx.fit.LogExpoCurve2'>,
 'Exponential (log x) (3)': <class 'qubx.fit.LogExpoCurve3'>,
...

Fitters

Value:

{'Levenburg-Marquardt': <class 'qubx.fit.LMFitter'>,
 'Simplex': <class 'qubx.fit.SimplexFitter'>,
 'Simplex-LM': <class 'qubx.fit.Simplex_LM_Fitter'>}

GaussianCurves

Value:

[<class 'qubx.fit.GaussianCurve'>,
 <class 'qubx.fit.DblGaussianCurve'>,
 <class 'qubx.fit.TriGaussianCurve'>,
 <class 'qubx.fit.QuadGaussianCurve'>,
 <class 'qubx.fit.QuintGaussianCurve'>]

GaussianUnitCurves

Value:

[<class 'qubx.fit.GaussianUnitCurve'>,
 <class 'qubx.fit.DblGaussianUnitCurve'>,
 <class 'qubx.fit.TriGaussianUnitCurve'>,
 <class 'qubx.fit.QuadGaussianUnitCurve'>,
 <class 'qubx.fit.QuintGaussianUnitCurve'>]