The Semi-parametric Mixed Effect Model Of Spatial Correlation And Its Application

As the information age,the field involved in the area is increasing,and the included information characteristics are also gradually increased.The periodic feature and spatial association feature of the data often have an important impact in practical applications,so in the modeling process.It is important to consider these information characteristics,but the traditional semi-parameter model mixed effect model can only solve the internal dependence characteristics of the vertical data,and cannot seize the spatial correlation between the individual subjects,so this article is based on the basis of this model.The improvement is improved,thereby increasing the utilization of data and the universality.First,this paper proposes a single point analysis model on the basis of a semi-parameter model.The model passes through the introduction of the cycle function,thereby better grasping the periodic feature of the data,while combining the selection of the model by point by point by point,thus Improve the fitting and prediction accuracy of the data;secondly,the spatial correlation is considering the spatial correlation on the basis of the half-parameter mixed effect model,and the distance and position-related parameters are introduced by the covariance of the random effect,thereby constructing different test individuals.Spatial correlation between;The relevant nature is theoretical.Finally,the simulation experiment is performed on the semi-parametric hybrid effect model considering spatial relevance,and the proposed predictive effect is obtained;and the proposed single-point analysis model is applied to the subgrade settlement prediction of the Qinghai-Tibet Railway,the frozen soil 15 test points in 2010 to the 117-month road base highprofile difference in 2019,and compared with a wide gray GM(1,1)model applied in this area,further prove the effectiveness and practical use of the model.