New Perspectives on Centering

Data matrix centering is an ever-present yet under-examined aspect of data analysis. Functional analysis (FDA) often operates with a default such that the vectors in one dimension have mean zero. We find along other identifies novel useful mode variation beyond those familiar FDA. explore ambiguities both orientation and nomenclature. Differences between centerings their potential interaction can be easily misunderstood. propose unified framework new terminology for operations. clearly demonstrate intuition behind consequences each choice informative graphics. also direction energy hypothesis test as part series diagnostics determining which best set. application these several FDA settings.