Consistency of overlapping batch variances

Overlapping batch statistics estimate the variance of point estimators using overlapping batches (Schmeiser, Avramidis and Hashem 1990). In this paper we study sufficient conditions for the mean-squared-error (mse) consistency of overlapping batch variances (OBV). We show that both the bias and the variance of OBV go to zero as the sample size increases under very mild conditions. In fact we conjecture that whenever overlapping batch means (OBM) is consistent OBV is consistent except for symmetric Bernoulli data. We furthermore conjecture that overlapping batch statistics is mse consistent as the estimator of the variance of all mean-like estimators.

• Type: Conference
• Year: 1994
• Where: WSC
• Authors: Demet C. Wood, Bruce W. Schmeiser
• Keywords: Batch Statistics, Batch Variances, Modeling And Simulation, Variance, WSC 1994