A0672
Title: Specification testing with mixed discrete and continuous predictors: A projection-based adaptive-to-model approach
Authors: Xuehu Zhu - Xi'an Jiaotong university (China) [presenting]
Abstract: A nonparametric projection-based adaptive-to-model specification test is proposed, which can be used to check the regression model with both discrete and continuous predictors. Unlike existing locally smoothing kernel tests, the new test behaves like a locally smoothing test as if the number of continuous predictors was one, and it can detect the local alternative hypotheses distinct from the null hypothesis at the rate that is only related to one continuous predictors. The test statistic is asymptotically normal under the null hypothesis such that critical values are easily determined. Simulation studies and real data analysis are then conducted to illustrate the performance of the new test and compare it with existing tests.
