摘要：AbstractEstimation of time delays from a noisy and gapped data is one of the simplest data analysis problems in astronomy by its formulation. But as history of real experiments show, the work with observed datasets can be quite complex and evolved. By analyzing in detail previous attempts to build delay estimation algorithms we try to develop an automatic and robust procedure to perform the task. To evaluate and compare different variants of the algorithms we use real observed datasets which have been objects of past controversies. In this way we hope to select the methods and procedures which have highest probability to succeed in complex situations. As a result of our investigations, we propose an estimation procedure which can be used as a method of choice in large photometric experiments. We cannot claim that the proposed methodology works with any reasonably well sampled input dataset. However we hope that the steps taken are in correct direction and the developed software will be useful for observational astronomers.
