本文借助经验欧氏似然构造一类新的拟合优度检验,并讨论在简单零假设下,检验统计量的极限分布;然后利用极大似然估计所得的参数估计量构造复合零假设下的检验函数,并讨论其极限性质;最后把所得的检验与已有的Pearsonχ2检验和KS检验进行模拟比较,模拟结果显示基于经验欧氏似然的检验比其他的检验有相对的优势:功效较高、计算简单等,在应用上更具有推广价值。
A goodness of fit test is constructed by means of empirical Euclidean likelihood,and the limit null distribution of the test statistics is derived.Then,the parameters are estimated by maximum likelihood estimate(MLE) under the composite null hypothesis.Subsequently,the corresponding test statistics is constructed and its limit null distribution is given.Finally,the new test is compared with chi-square test and Kolmogorov-Smirnov(KS) test by simulations.Simulation results show that the new test is more powerful and has smaller computational complexity.Therefore,the new test shows favorable application and potential prospect.