My Math Forum Is there an analytical solution for a regression of this form?

 May 2nd, 2014, 06:58 AM #1 Newbie   Joined: May 2014 From: Manhatten Posts: 1 Thanks: 0 Is there an analytical solution for a regression of this form? Hi guys, I know I can run a quadratic regression and get an analytical solution if my function looks like this: $\displaystyle y = f(x) = a*x^2 + b*x + c$ However, the problem I'm working on involves following functional relationship: $\displaystyle y = f(x) = (a*x^2+b*x)/(c*x+d)$ Is it still possible to get an analytical solution out of this? Possibly also with quatratic/polynomial regression?
 May 2nd, 2014, 01:03 PM #2 Global Moderator   Joined: May 2007 Posts: 6,305 Thanks: 525 Your main problem would be if cx+d = 0 for some x in the domain.
 May 8th, 2014, 01:15 AM #3 Senior Member   Joined: Aug 2011 Posts: 333 Thanks: 8 The fitting of this pattern of function is possible if you consider the equation : a*x²+b*x-c*y*x-d*y = 0 Of course, one of the four parameters must be arbitrary fixed and the regression will be carried out for three parameters only (because proportionality). This is a kind of problem of "equation fitting" or of "implicit function" fitting. For example, see Section 5 in the paper : Régressions coniques, quadriques, circulaire, sphérique (written in French, but the equations are undertandable in any language)

 Tags analytical, form, regression, solution

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post munkifisht Calculus 2 June 13th, 2012 12:02 PM sciacallojo2 Economics 7 August 9th, 2011 11:37 AM jangolobow Calculus 1 April 2nd, 2010 08:51 AM ask2 Calculus 2 November 13th, 2009 02:08 AM handy11 Algebra 0 December 31st, 1969 04:00 PM

 Contact - Home - Forums - Cryptocurrency Forum - Top