Elastic Functional Regression¶
These functions compute elastic standard, logistic, and m-logistic regression models. This code is experimental and results are not guaranteed
Regression Models and Prediction¶
-
elastic_regression
(f, y, timet; B=None, lambda=0, df=20, max_itr=20, smooth=false)¶ Calculate elastic regression from function data f, for response y
f
array (M,N) of N functionsy
vector (N) of responsestimet
vector (N) describing time samplesB
matrix describing basis functions (M,N) (default=None generates a B-spline basislambda
regularization parameterdf
degree of freedom of basismax_itr
maximum number of iterationssmooth
smooth data
Returns Dict describing regression:
alpha
interceptbeta
regression functionfn
aligned functionsqn
aligned srsfsgamma
warping functionsq
original srsfsB
basis functionstype
type of regressionb
coefficientsSSE
sum of squared error
-
elastic_logistic
(f, y, timet; B=None, df=20, max_itr=20, smooth=false)¶ Calculate elastic logistic regression from function data f, for response y
f
array (M,N) of N functionsy
vector (N) of responsestimet
vector (N) describing time samplesB
matrix describing basis functions (M,N) (default=None generates a B-spline basisdf
degree of freedom of basismax_itr
maximum number of iterationssmooth
smooth data
Returns Dict describing regression:
alpha
interceptbeta
regression functionfn
aligned functionsqn
aligned srsfsgamma
warping functionsq
original srsfsB
basis functionstype
type of regressionb
coefficientsLL
logistic loss
-
elastic_mlogistic
(f, y, timet; B=None, df=20, max_itr=20, smooth=false)¶ Calculate elastic m-logistic regression from function data f, for response y
- ``f: array (M,N) of N functions
- ``y: vector (N) of responses
- ``timet: vector (N) describing time samples
- ``B: matrix describing basis functions (M,N) (default=None generates a B-spline basis
- ``df: degree of freedom of basis
- ``max_itr: maximum number of iterations
- ``smooth: smooth data
Returns Dict describing regression:
alpha
interceptbeta
regression functionfn
aligned functionsqn
aligned srsfsgamma
warping functionsq
original srsfsB
basis functionstype
type of regressionb
coefficientsn_classes
number of classesLL
logistic loss
-
elastic_prediction
(f, timet, model; y=None, smooth=false)¶ Prediction from elastic regression model
f
functions to predicttimet
vector describing time samplesmodel
calculated model (regression, logistic, mlogistic)y
true responses (default = None)smooth
smooth data (default = false)
Returns:
y_pred
predicted valuey_labels
labels of predicted valuePerf
Performance metric if truth is supplied