"利用鞅差相关系数检验参数回归模型异方差性问题的研究"

is test's consistency under the alternative hypothesis. Because the asymptotic null distribution of the test statistic depends on the unknown distribution of the data, we employ the wild bootstrap method to approximate this null distribution. Furthermore, through Monte Carlo simulation, we compare the performance of our test with existing methods in finite sample situations, demonstrating its relatively higher power at a given level of significance. Keywords: Martingale Difference Divergence; parametric regression model; heteroscedasticity. This article introduces a new measurement of conditional uncorrelatedness, known as the Martingale Difference Divergence, and utilizes this measure to develop a test for heteroscedasticity in parametric regression models. The author derives the asymptotic null distribution of this test statistic and proves its consistency under the alternative hypothesis. As the asymptotic null distribution of the test statistic is dependent on the unknown distribution of the data, the author employs the wild bootstrap method to approximate this distribution. Additionally, through Monte Carlo simulation, the author compares the performance of this test with existing methods in situations of limited sample sizes, demonstrating its relatively higher power at a given level of significance.