DATA REQUIREMENTS IN STATISTICAL DECISION-SUPPORT SYSTEMS - FORMULATION AND SOME RESULTS IN CHOOSING SUMMARIES

The problem of determining data requirements in cases where statistica l query answers are desired is studied. Specifically, we consider the value of storing aggregate data that can be used to speed up answering such queries, but at the potential costs of incomplete information du e to either estimation error or staleness, as well as increased costs of update. We formulate the overall optimization problem for design, a nd decompose it into several subproblems that can be separately addres sed. Two of these subproblems are the choice of update method, and cho ice of aggregates. Qualitative results are given regarding the selecti on of update policy, and design heuristics, based on numerical experim ents, are given for single-attribute Legendre polynomial aggregates. M ultivariate Legendre aggregates are also discussed, and suggestions fo r future research are given.