File talk:AsymptoticExpansionE1.png
Dear all! I would like to improve this graph, because I think it is wrong: relative errors should be positive for N=1 (red), N=3 (yellow) and N=5 (pink) and negative for N=2 (green) and N=4 (blue). I agree completely with definitions of exact function (E1(z)), approximate functions, for different values of N (E1(z, N)) and relative error of the latter with respect to the former ((E1(z, N)-E1(z))/E1(z)), so that is not the cause ... Also I disagree at which points plots for N=m and N=m+2 cross each other (end also crossing of N=1 with N=5); I think that that should occur when these two approximations match each other, so "crossing" of N=1 with N=3 should occur at x=2 (relative error approx. +38.378%), N=2 with N=4 at x=3 (relative error approx. -15.209%), N=3 with N=5 at x=4 (relative error approx. +6.011%) and N=1 with N=5 approximately at x=2.84978 (relative error approx. +28.416%). — Preceding unsigned comment added by THT16 (talk • contribs) 11:01, 28 May 2017 (UTC)
