Numerical Methods.

Computing with real numbers on a machine that cannot hold them: floating point, error, solvers and the conditions under which they converge. Covers 8 topics.

• Numerical Methods · 1 of 3

  Linear Systems

  What defines a singular square matrix?

  AIts diagonal is 0BIts determinant is 1CIt has no inverseDIt is not symmetric

• Numerical Methods · 2 of 3

  Interpolation & Fitting

  What determines if extrapolating a fitted curve beyond its data gives a trustworthy prediction?

  AA correct model formBThe R-squared of the fitCThe spacing of the dataDThe count of points

• Numerical Methods · 3 of 3

  Floating Point Representation

  Does compensated summation remove all rounding error?

  ANo, it shrinks the error but not to zeroBYes, it removes rounding error fullyCNo, it only halves the error of plain summationDYes, provided the terms are sorted first