math/

jblinfit.pro

top source jblinfit

Math

result = jblinfit(x, y, dimension=dimension, chisqr=chisqr, covar=covar, measure_errors=measure_errors, prob=prob, sigma=sigma, yfit=yfit)

Performs least squares fitting to a straight line, but can perform it over one particular dimension for a multi-dimensional data set. Acts like the built-in function LINFIT if DIMENSION is not set.

Return value

Returns the parameters of the linear fit. If X has dimensions [D1, D2, D3... DM], and DIMENSION=N, then Result has dimensions [2, D1...DN-1, DN+1...DM]. Result[0,....] contains the constant term and Result[1,...] contains the slope.

Parameters

x in: Array containing the independent variable values.
y in: Array containing the dependent variable values.

Keywords

dimension in: If X and Y are multi-dimensional arrays, perform the fitting accross this dimension.
chisqr in: Output variable containing the chi squared values.
covar in: Output variable containing the covariance matrix.
measure_errors in: Array containing the measurement errors in Y.
prob in: Output variable containing the probability of obtaining a fit with at least this chi squared value.
sigma in: Output variable containing uncertainties in fit parameters.
yfit in: Output variable containing the values of the dependent variable at the X locations, according to the fit.

Examples

    seed = 43l
    n = 10000L
    x = 1e-3 * findgen(n)
    y1 = x + randomn(seed, n)
    y2 = 10. - 2 * x + 0.5 * randomn(seed, n)
    x = [[x],[x]]
    y = [[y1],[y2]]

    result = jblinfit(x, y, sigma=sigma, yfit=yfit, dimension=1)

    xax = [0,10]
    cgplot, psym=3, x, y1, yrange=minmax(y), color='red'
    cgplot, /overplot, psym=3, x, y2, color='blue'
    cgplot, /overplot, xax, result[0,0] + result[1,0]*xax, color='red'
    cgplot, /overplot, xax, result[0,1] + result[1,1]*xax, color='blue'

    cgtext, x[9000], yfit[9000,0] + 2, color='red', align=1, $
      string(result[0,0], sigma[0,0], result[1,0], sigma[1,0], $
      format='(%"y = (%0.2f +/- %0.2f) + (%0.3f +/- %0.3f) x")'), $
      charsize=1.5
    cgtext, x[9000], yfit[9000,1] - 3, color='blue', align=1, $
      string(result[0,1], sigma[0,1], result[1,1], sigma[1,1], $
      format='(%"y = (%0.2f +/- %0.2f) + (%0.3f +/- %0.3f) x")'), $
      charsize=1.5

Author information

History:: Written by: Jeremy Bailin 28 March 2011 Initial writing.

File attributes

Modification date:	Mon Mar 28 16:39:22 2011
Lines:	136

Documentation for jbiu/

Generated by IDLdoc