QUASI-LIKELIHOOD METHOD IN ESTIMATION OF SPATIAL DEPENDENCE

摘要

The present invention include step estimate a doctor term square differences variogram model likelihood function Q (β) parameter beta be related to doctor's possibility use it is hollow it is m- rely on model estimation to β ' to the maximum extent, it is obtained from quasi- likelihood because from estimation formula including obtaining close β ' an estimations formula by an iterative process, not only with the excellent statistical characteristics as aspect to the different dependence models with different parameter values, it is compared with the traditional method, the inertia group that you can not rule out the upcoming haneundeseo of randomness is the special lag order of preassignment one, the parameter of estimation variogram model is without any process to be averaged, therefore effective, it has the efficiency loss of utilization benefit reality being avoided that in averaging process.