TESTING FOR STRUCTURAL CHANGE IN TIME-VARYING NONPARAMETRIC REGRESSION MODELS
针对时变非参数回归模型，提出一个基于核的L2检验统计量，用于检验回归函数在给定时间段内是否形状不变，并推导了渐近分布和自助法性质。
In this paper, we consider a nonparametric model with a time-varying regression function and locally stationary regressors. We are interested in the question whether the regression function has the same shape over a given time span. To tackle this testing problem, we propose a kernel-based L 2 -test statistic. We derive the asymptotic distribution of the statistic both under the null and under fixed and local alternatives. To improve the small sample behavior of the test, we set up a wild bootstrap procedure and derive the asymptotic properties thereof. The theoretical analysis of the paper is complemented by a simulation study and a real data example.