1 / 8

Practical Issues

Practical Issues. Stephen Mildenhall CNA Re. FFT Roadmap. exponentiation. Addition Exp(x+y) Easy?. Multiplication Exp(x)Exp(y) Hard?. logarithm. FFT Roadmap. Random Variable: X Fourier Transform: t  E(exp(itX)). Random Variables X, Y X + Y. t  E(exp(itX)) t  E(exp(itY))

wei
Download Presentation

Practical Issues

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Practical Issues Stephen MildenhallCNA Re

  2. FFT Roadmap exponentiation Addition Exp(x+y) Easy? Multiplication Exp(x)Exp(y) Hard? logarithm

  3. FFT Roadmap Random Variable: XFourier Transform: t  E(exp(itX)) Random Variables X, Y X + Y t  E(exp(itX)) t  E(exp(itY)) tE(exp(itY)).E(exp(itY)) MGF or Fourier Transform Same information

  4. FFT Roadmap • Fourier Transform is a continuous function • Discrete Fourier Transform (DFT) is a sample of the Fourier Transform • Contains approximately the same information • Two-way street, convert X to DFT(X) to X • Computes sums (convolutions) • DFT is just matrix multiplication DFT(X)= M X • Fast Fourier Transform is a very, very fast way of computing DFT(X) • Available in MALT, Excel, Matlab, SAS, C/C++ • Black-box or learn to program

  5. FFT Roadmap DFT(X), DFT(Y)(n x 1 vectors of complex numbers, computed very quickly using FFT) Random variables X, Y (n x 1 vectors of real numbers, sum=1) DFT(X) x DFT(Y)(component-by-component product) Inverse DFT = distribution of X+Y (!)

  6. n X Y X+Y 0 0.2 0 0 1 0.2 1 0.2 2 0.2 0 0.2 3 0.2 0 0.2 4 0.2 0 0.2 5 0 0 0.2 6 0 0 0 7 0 0 0 n Re(FTX) Im(FTX) Re(FTY) Im(FTY) Re(FTX+Y) Im(FTX+Y) 0 1.0000 0.0000 1.0000 0.0000 1.0000 0.0000 1 0.0000 -0.4828 0.7071 -0.7071 -0.3414 -0.3414 2 0.2000 0.0000 0.0000 -1.0000 0.0000 -0.2000 3 0.0000 -0.0828 -0.7071 -0.7071 -0.0586 0.0586 4 0.2000 0.0000 -1.0000 0.0000 -0.2000 0.0000 3 0.0000 0.0828 -0.7071 0.7071 -0.0586 -0.0586 2 0.2000 0.0000 0.0000 1.0000 0.0000 0.2000 1 0.0000 0.4828 0.7071 0.7071 -0.3414 0.3414 Step-by-Step Convolution Calculation Distributions FFTs

  7. MALT Site • http://www.mynl.com/MALT/home.html • Written in response to COOS Call Program, 2000-01 • Windows 2000/NT, IE 5 required • Download and install COM objects required • Can script or use in Excel/VBA

  8. Independent Clayton Gumbel Venter HRT Warning to ERM Practice • A correlation matrix is not enough to describe the correlations at the tails

More Related