Hierarchical modelling of archaeomagnetic data and curve estimation by moving average technique

摘要: A Bayesian hierarchical modelling is proposed for the different sources of scatter occurring in archaeomagnetism, which follows natural sampling process implemented by laboratories field. comparison made with stratified statistics commonly used up to now. The corrects disturbance resulting from variability number specimens taken each sample or site. There no need publish results at level if a descending hierarchy verified. In this case, often verified archaeomagnetic data, only site are useful geomagnetic reference curve building. Typically, study least 20 samples will give an α95i 5 per cent close optimal fixed mi and errors random zero mean (no systematic errors). precision on itself essentially controlled, through elliptic statistics, points window dating errors, rather than confidence angles α95ij (if hierarchy). distribution reveals influence width. moving average technique well adapted numerous very dated data evenly distributed along time. It not global functional approach, but (linear) local one.