Local disclosure control techniques of statistical microdata sets

在这一年里，我们在统计微观数据集的披露控制技术方面取得了一些重要的成果。Takemura在国际统计研究所会议上就《基于偶发表元概率交互作用的加性建模的每记录识别风险评估》一文作了特邀演讲。这篇论文将包括在国际调查统计学家协会出版的另一份会议记录中。《日本统计学会杂志》将发表一篇论文《微观数据集中单个记录的披露风险评估的最小不安全和最大安全变量集》。星野信明的论文《关于皮特曼抽样公式在微观数据披露风险评估中的应用》发表在《官方统计杂志》上。作为一项新的研究成果，卡内基梅隆大学发表了两篇讨论论文《关于条件逆高斯-泊松分布的随机聚类》和《关于条件逆高斯-泊松分布的极限随机划分结构》。

We obtained several important results on the disclosure control techniques of statistical microdata sets in this year.Takemura gave an invited talk on the paper "Evaluation of per-record identification risk by additive modeling of interaction for contingency table cell probabilities" at the meeting of International Statistical Institute. This paper will be included in a separate proceedings published by International Association of Survey Statisticians. A paper "Minimum unsafe and maximum safe sets of variables for disclosure risk assessment of individual records in a microdata set" will be published in Journal of Japan Statistical Society The paper "On application of Pitman's sampling formula to microdata disclosure risk assessment" by Nobuaki Hoshino was published in Journal of Official Statistics. As a new research results, two discussion papers "On the random clustering with the conditional inverse Gaussian-Poisson distribution" and "On limiting random partition structure derived from the conditional inverse Gaussian-Poisson distribution" are published as discussion paper in Carnegie Mellon university.