zbMATH — the first resource for mathematics

Testing the equality of nonparametric regression curves. (English) Zbl 0771.62034
Summary: This paper proposes a test for the equality of nonparametric regression curves that does not depend on the choice of a smoothing number. The test statistic resembles in spirit the Kolmogorov-Smirnov statistic and it is easy to compute. It is powerful under alternatives that converge to the null hypothesis at a rate \(n^{-1/2}\). The disturbance distributions are arbitrary and possibly unequal, and conditions on the regressors distribution are very mild. A Monte Carlo study illustrates the performance of the test in small and moderate samples. We also study extensions to multiple regression, and test the equality of several regression curves.

62G10 Nonparametric hypothesis testing
Full Text: DOI
[1] Billingsley, P., Convergence of probability measures, (1968), Wiley New York · Zbl 0172.21201
[2] Barry, D.; Hartigan, J.A., An omnibus test for departures from constant Mean, Ann. statist., 18, 1340-1357, (1990) · Zbl 0706.62046
[3] Hall, P.; Hart, J.D., Bootstrap test for difference between means in nonparametric regression, J. amer. statist. assoc., 85, 1039-1049, (1990) · Zbl 0717.62037
[4] Härdle, W.; Marron, J.S., Semiparametric comparison of regression curves, Ann. statist., 18, 63-89, (1990) · Zbl 0703.62053
[5] Hong-zhi, A.; Bing, C., A Kolmogorov-Smirnov type statistic with application to test for nonlinearity in time series, Internat. statist. rev., 59, 287-307, (1991) · Zbl 0748.62049
[6] King, E.C., A test for the equality of two regression curves based on kernel smoothers, ()
[7] King, E.C.; Hart, J.D.; Wehrly, T.E., Testing the equality of two regression curves using linear smoothers, Statist. probab. lett., 12, 239-247, (1991)
[8] Rice, J., Bandwidth choice for nonparametric regression, Ann. statist., 12, 1215-1230, (1984) · Zbl 0554.62035
[9] Shorack, G.R.; Wellner, J.A., Empirical processes with applications to statistics, (1986), Wiley New York · Zbl 1170.62365
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.