Dear All,
I want to run the non-linear estimation below. My sample has several periods (0 < t< T2) and I want to estimate the non-linear equations with the same coefficients. Do you know how I can estimate the system with nlsur, so that the coefficients {a}, {b} are estimated taking into account the information in both time periods? One issue I have is that in some settings the first time period might be too short to estimate the first equation separately, but I do not want to drop the information carried in these data points.
Period of First Generation (0 < t< T1):
D.x1 = ({a} + {b}*x1/{N})(N - x1)
Period of Second Generation (T1 < t < T2):
D.x2 = ({a} + {b}*(x1+x2)/{N2})({N2} - (x1+x2)) + {a2}*({a3} + {b1}*x2/{N2})*x1
D.x1 = (1 - {a2})({a} + {b}*(x1+x2)/{N2})({N2} - (x1+x2)) - {a2}*({a3} + {b1}*x2/{N2})*x1
Thanks,
Socrates
I want to run the non-linear estimation below. My sample has several periods (0 < t< T2) and I want to estimate the non-linear equations with the same coefficients. Do you know how I can estimate the system with nlsur, so that the coefficients {a}, {b} are estimated taking into account the information in both time periods? One issue I have is that in some settings the first time period might be too short to estimate the first equation separately, but I do not want to drop the information carried in these data points.
Period of First Generation (0 < t< T1):
D.x1 = ({a} + {b}*x1/{N})(N - x1)
Period of Second Generation (T1 < t < T2):
D.x2 = ({a} + {b}*(x1+x2)/{N2})({N2} - (x1+x2)) + {a2}*({a3} + {b1}*x2/{N2})*x1
D.x1 = (1 - {a2})({a} + {b}*(x1+x2)/{N2})({N2} - (x1+x2)) - {a2}*({a3} + {b1}*x2/{N2})*x1
Thanks,
Socrates