Data-Driven Version of Neyman's Smooth Test of Fit

摘要: Abstract Neyman's smooth test for testing uniformity is a recognized goodness-of-fit procedure. As stated by LaRiccia, the can be viewed as compromise between omnibus procedures, with generally low power in all directions, and procedures whose focused direction of specific alternative. The basic idea behind this to embed null density into, say, k-dimensional exponential family then construct an asymptotically optimal parametric problem. resulting procedure k components. most difficult problem related using choice k. Recommendations statistical literature are sometimes confusing. Some authors advocate small number components, whereas others show that some situations larger components profitable. All existing suggestions concerning how select exploit fact preliminary knowledge about possible In article, new data-driven met...