[R] Regression with no-intercept

From: Alexandra R. M. de Almeida <alexandrarma_at_yahoo.com.br>
Date: Wed 18 Jan 2006 - 05:24:24 EST


Dear R users,  

 There is a method called "style analysis" where you make a regression being Y=fund yield and X=benchmarks yield, where we have the restrictions to calculatethe linear regression:  

  1. The regression must don have the intercept term.
  2. The coefficient sum must be one.
  3. All coefficients must be upper than zero.

 These restrictions are necessary to give to regression the sense of percentual benchmarks contributtions.   

 Of course the least square method can't do that. So I used the Quadratic Programation method.   

 The question is: How I don't use the least square method, how can I evaluate (without graphics) the good/bad fit?  

 Another: How can I interpret the exit of the intercept term in the regression?  

 Thanks   

  Alexandra  

  Alexandra R. Mendes de Almeida                                                                   


        [[alternative HTML version deleted]]



R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html Received on Wed Jan 18 05:47:07 2006

This archive was generated by hypermail 2.1.8 : Wed 18 Jan 2006 - 14:06:59 EST