[EMAIL PROTECTED] (Mohammad Ehsanul Karim) wrote in message news:<[EMAIL PROTECTED]>... > [EMAIL PROTECTED] (David Reilly) wrote in message news:<[EMAIL PROTECTED]>... > > > Not discussed BY YOU is the assumption that the parameters are > > constant over the range of observation . > > > > Will you be kind enough to give me a bit more hint about this > assumption ? My references can't provide me more help!
http://www.princeton.edu/~gchow/ details Prof. Chow's work . Specifically he developed tests for whether coefficients from 1 set of data were statistically different from coefficients from a second set . What we have done in AUTOBOX is search for that point in time that generates the most significant difference between two groups and then evaluating whether this is statistically significant. If it is then one can conclude that the model parameters have changed from one regime ( prior to time point "t" ) to the next regime. hope this helps ... dave reilly automatic forecasting systems http://www.autobox.com > > > Gregory Chow's work on testing the hypothesis of a constant set of > > parameters > > underpins some of the work that we do in AUTOBOX to test for breal > > points. > > > > http://www.autobox.com > > > > regards > > > > dave reilly > > AFS > _______________________ > > Mohammad Ehsanul Karim <[EMAIL PROTECTED]> > Institute of Statistical Research and Training > University of Dhaka, Dhaka- 1000, Bangladesh > _______________________ . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
