METHOD FOR STATISTICAL DISCLOSURE LIMITATION

摘要

A kind of method and system, the absolute or continuous micro data in open limitation (SDL) for guaranteeing statistics, while keeping analytical quality microdata. New methodology is using similitude (1) sampling (rather than census) and some adjustment between SDL, including imputes to, drain message, (2) subset is discharged, rather than original data set and some adjustment are put on record still in disclosure risk. Tentative sampling reduces monetary cost and census, but needs some information losses. Similarly, release subset reduces open cost and full database, but needs some information losses. Therefore, the optimum test method of sampling can be used for counting open limitation. This method is that database is divided into risk stratum, and optimal probability replaces, optimal probability double sampling, optimal sampling weight-calibration.