version 1.0.0.0 (5.93 KB) by
Leonidas Bantis

Fits the so called restricted cubic spline via least squares and obtains 95% bootstrap based CIs.

%Fits the so called restricted cubic spline via least squares (see Harrell

%(2001)). The obtained spline is linear beyond the first and the last

%knot. The truncated power basis representation is used. That is, the

%fitted spline is of the form:

%f(x)=b0+b1*x+b2*(x-t1)^3*(x>t1)+b3*(x-t2)^3*(x>t2)+...

%where t1 t2,... are the desired knots.

%95% confidence intervals are provided based on the bootstrap procedure.

%For more information see also:

%Frank E Harrell Jr, Regression Modelling Strategies (With application to

%linear models, logistic regression and survival analysis), 2001,

%Springer Series in Statistics, pages 20-21.

%

%INPUT ARGUMENTS:

%x: A vector containing the covariate values x.

%y: A vector of length(x) that contains the response values y.

%knots: A vector of points at which the knots are to be placed.

% Alternatively, it can be set as 'prc3', 'prc4', ..., 'prc8' and 3

% or 4 or...8 knots placed at equally spaced percentiles will be used.

% It can also be set to 'eq3', 'eq4', ...,'eq8' to use 3 or 4 or ...

% or 8 equally spaced knots. There is a difference in using one of

% these strings to define the knots instead of passing them directly

% as a vector of numbers and the difference involves only the

% bootstrap option and not the fit itself. When the bootstrap is used

% and the knots are passed in as numbers, then the knot sequence will

% be considered fixed as provided by the user for each bootstrap

% iteration. If a string as the ones mentioned above is used, then

% the knot sequence is re-evaluated for each bootstrap sample

% based on this choice.

%

%OPTIONAL INPUT ARGUMENTS: (These can be not reached at all or set as [] to

%proceed to the next optional input argument):

%

%bootsams: The number of bootstrap samples if the user wants to derive 95% CIs.

%atwhich: a vector of x values at which the CIs of the f(x) are to be evaluated.

%plots: If set to 1, it returns a plot of the spline and the data.

% Otherwise it is ignored. This input argument can also not be reached

% at all. (It also plots the CIs provided that they are requested).

%

%OUTPUT ARGUMENTS:

%bhat: the estimated spline coefficients.

%f: a function handle from which you can evaluate the spline value at a

% given x (which can be a scalar or a vector). For example ff(2) will

% yield the spline value for x=2. You can use a vector (grid) of x values to

% plot the f(x) by requesting plot(x,f(x)).

%sse: equals to sum((y-ff(x)).^2)

%knots: the knots used for fitting the spline.

%CI : 95% bootstrap based confidence intervals.

% Obtained only if the bootstrap is requested and and only fot the

% points at which the CIs were requested. Hence, CI is a three column

% matrix with its first column be the spline value at the points

% supplied by the user, and the second and third column are

% respectively the lower and upper CI limits for that points.

%

%References: Frank E. Harrell, Jr. Regression Modeling Strategies (With

%applications to linear models, logistic regression, and survival

%analysis). Springer 2001.

%

%

%Code author: Leonidas E. Bantis,

%Dept. of Statistics & Actuarial-Financial Mathematics, School of Sciences

%University of the Aegean, Samos Island.

%

%E-mail: leobantis@gmail.com

%Date: January 14th, 2013.

%Version: 1.

Created with
R2010a

Compatible with any release

**Inspired:**
Covariate-Adjusted Restricted Cubic Spline Regression

Create scripts with code, output, and formatted text in a single executable document.