การกำหนดจำนวนกลุ่มตัวอย่างสำหรับการวิเคราะห์ความถดถอยโพรบิทแบบเกาซ์เซียนคอพพูลา / กรกฎ วัฒนวีร์ = Determination of number of sample groups for probit regression analysis with Gaussian copula / Korrakod Vadhanavira
To determine the number of sample groups for probit regression analysis with Gaussian copula, in order to control the precision and the reliability of the estimated parameter, by employing the sample size determination methodology in the sampling theory. The procedure is validated by comparing the error of the parameter estimation from simulated data to the level of precision and reliability suggested by the sample size determination methodology. We found that the estimation errors of β⁺ₒ σ²[subscript β[subscript g]] and [rho] tend to follow the precision and reliability under the sample survey theory when group sample size, the sample size in each group, is higher. We also provide adequate group sample sizes that insure the validity of the method in some specific cases.