I-spline

In the mathematical subfield of numerical analysis, an I-spline[1][2] is a monotone spline function.

An I-spline family of order three with four interior knots.

Definition

A family of I-spline functions of degree k with n free parameters is defined in terms of the M-splines Mi(x|k, t)


I_i(x|k,t) = \int_L^x M_i(u|k,t)du,

where L is the lower limit of the domain of the splines.

Since M-splines are non-negative, I-splines are monotonically non-decreasing.

Computation

Let j be the index such that tj  x < tj+1. Then Ii(x|k, t) is zero if i > j, and equals one if j  k + 1 > i. Otherwise,


I_i(x|k,t) = \sum_{m=i}^j (t_{m+k+1}-t_m)M_m(x|k+1,t)/(k+1).

Applications

I-splines can be used as basis splines for regression analysis and data transformation when monotonicity is desired (constraining the regression coefficients to be non-negative for a non-decreasing fit, and non-positive for a non-increasing fit).

References

  1. Curry, H.B.; Schoenberg, I.J. (1966). "On Polya frequency functions. IV. The fundamental spline functions and their limits". J. Analyse Math. 17: 71107. doi:10.1007/BF02788653.
  2. Ramsay, J.O. (1988). "Monotone Regression Splines in Action". Statistical Science 3 (4): 425441. doi:10.1214/ss/1177012761. JSTOR 2245395.


This article is issued from Wikipedia - version of the Thursday, October 15, 2015. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.