14/3/08

statistics software

http://www.stat.berkeley.edu/~stark/Code/index.htm

Software by P.B. Stark and collaborators

  • SticiGui interactive statistics textbook
  • Java applets for statistics
  • bvls.f is a FORTRAN subroutine to solve least-squares problems with bounds on the variables. It calls qr.f. See Stark, P.B., and R.L. Parker, 1995. Bounded-variable least-squares: an algorithm and applications, Comp. Stat., 10, 129-141.
  • qr.f is a FORTRAN subroutine to compute QR decompositions in a stable way; it is called by bvls.f
  • sbl1.f is a FORTRAN program to find bounds on linear functionals of an n-vector subject to an l1 constraint on the misfit to a set of linear relations, and linear inequality constraints on the variables. It calls bvls.f and qr.f.
  • sbvq.f is a FORTRAN program to find bounds on linear functionals of an n-vector subject to a quadratic constraint on the misfit to a set of linear relations, and linear inequality constraints on the variables. It calls bvls.f and qr.f.
  • sci is an S+ function (to be loaded into S+ using "get") to compute nonequivariant simultaneous confidence intervals for the components of the mean of a multivariate normal, in a way that the intervals are less likely to contain zero than traditional methods. See Benjamini, Y. and Stark, P.B., 1996. Nonequivariant Confidence Intervals Less Likely to Contain Zero, J. Am. Stat. Assoc., 91, 329-337.
  • Multitaper.zip is a compressed directory of MATLAB and C routines by I.K. Fodor to compute optimal tapers to estimate the power spectra of time series with gaps. See Fodor, I. and P.B. Stark, 2000. Multitaper Spectrum Estimation for Time Series with Gaps, IEEE Trans. Signal Processing, 48, 3472-3483.
  • LFA_Search is a collection of routines by Chad Schafer to find least-favorable alternatives and least-favorable prior probability distributions. The routines are designed to find minimax expected size confidence sets.

Last modified 1 February 2008. P.B. Stark. statistics.berkeley.edu/~stark/Code/index.htm

Sphere: Related Content

No hay comentarios: