Skip to content

Latest commit

 

History

History
21 lines (16 loc) · 1.45 KB

README.md

File metadata and controls

21 lines (16 loc) · 1.45 KB
Raw

splinefit

Estimate the relation between a scalar dependent variable Y and an N dimensional indepentent variable X using a spline based on data points. The spline tries to fit the data using a least squares approach. The continuity between the splines can be set seperately along each dimension. In each dimension, it is possible to set

  • no continuity (independent polynomials)
  • absolute value continuity (0th order), polynomials intersect along their mating borders
  • first derivative continuity (1st order), polynomials intersect and have same first derivative along their mating borders
  • second derivative continuity (2st order), polynomials intersect and have same first and second derivative along their mating borders
  • etc...

An N dimensional cartesian grid has to be defined, which does not have to be equidistant along any dimension. Each hypercube of this grid corresponds to an N dimensional polynomial. The polynomial structure that is used in all hypercubes can be defined by setting the maximum power along each dimension. A polynomial is then created that uses all possible mononomials upto these maximum powers.

See example/BasicsN2 to see an example of fitting a spline to data with 2 dimensional dependent variable X.

Dependencies

Only uses numpy and itertools

Available user functions

See docstrings for info.

  • model_from_data(X_fit, Y_fit, border_loc, poly_orders, deriv_orders)
  • MVSModel.evalSingle(self, x_point)
  • MVSModel.eval(self, x_points)