R - Faster Way to Calculate Rolling Statistics Over a Variable Interval

Question

I\'m curious if anyone out there can come up with a (faster) way to calculate rolling statistics (rolling mean, median, percentiles, etc.) over a variable interval of time (

Answer 1

Rcpp is a good approach if speed is your primary concern. I'll use the rolling mean statistic to explain by example.

Benchmarks: Rcpp versus R

x = sort(runif(25000,0,4*pi))
y = sin(x) + rnorm(length(x),0.5,0.5)
system.time( rollmean_r(x,y,xout=x,width=1.1) )   # ~60 seconds
system.time( rollmean_cpp(x,y,xout=x,width=1.1) ) # ~0.0007 seconds

Code for Rcpp and R function

cppFunction('
  NumericVector rollmean_cpp( NumericVector x, NumericVector y, 
                              NumericVector xout, double width) {
    double total=0;
    unsigned int n=x.size(), nout=xout.size(), i, ledge=0, redge=0;
    NumericVector out(nout);

    for( i=0; i width && ledge= (xout[i]-width) & x <= (xout[i]+width)
    out[i] = .Internal(mean( y[window] ))
  }
  return(out)
}

Now for an explantion of rollmean_cpp. x and y are the data. xout is a vector of points at which the rolling statistic is requested. width is the width*2 of the rolling window. Note that the indeces for the ends of sliding window are stored in ledge and redge. These are essentially pointers to the respective elements in x and y. These indeces could be very beneficial for calling other C++ functions (e.g., median and the like) that take a vector and starting and ending indeces as input.

For those who want a "verbose" version of rollmean_cpp for debugging (lengthy):

cppFunction('
  NumericVector rollmean_cpp( NumericVector x, NumericVector y, 
                              NumericVector xout, double width) {

    double total=0, oldtotal=0;
    unsigned int n=x.size(), nout=xout.size(), i, ledge=0, redge=0;
    NumericVector out(nout);


    for( i=0; i width && ledge