ビッグミクロデータの匿名性評価手法の開発

• Project/Area Number: 18H00835
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Review Section: Basic Section 07030:Economic statistics-related
• Research Institution: Kanazawa University
• Principal Investigator: 星野 伸明 金沢大学, 経済学経営学系, 教授 (00313627)
• Co-Investigator(Kenkyū-buntansha): 間野 修平 統計数理研究所, 数理・推論研究系, 教授 (20372948) ; 入江 薫 東京大学, 大学院経済学研究科(経済学部), 講師 (20789169) ; 佐井 至道 岡山商科大学, 経済学部, 教授 (30186910) ; 丸山 祐造 神戸大学, 経営学研究科, 教授 (30304728)
• Project Period (FY): 2018-04-01 – 2023-03-31
• Project Status: Granted (Fiscal Year 2022)
• Budget Amount: ¥16,900,000 (Direct Cost: ¥13,000,000、Indirect Cost: ¥3,900,000); Fiscal Year 2022: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000); Fiscal Year 2021: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000); Fiscal Year 2020: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000); Fiscal Year 2019: ¥3,250,000 (Direct Cost: ¥2,500,000、Indirect Cost: ¥750,000); Fiscal Year 2018: ¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
• Keywords: プライバシー / 統計的開示制限

Outline of Annual Research Achievements

昨年度に引き続き、復元抽出の族であるベル多項式分布の理論的解析を続けた。昨年度の研究で、連続な単体上の分布である正規化無限分解可能分布と、離散単体上の分布であるベル多項式分布の類似性について、二次モメントまでは比例することを明らかにした。今年度はより一般的に、ベル多項式分布が多項分布と正規化無限分解可能分布の混合で得られることを明らかにした。この結果から、二次モメントまでの比例も説明可能である。またこの結果は、ベル多項式分布を総度数で基準化した量の漸近分布が正規化無限分解可能分布になる事を意味する。この含意は、ベル多項式分布族の大標本理論の一般的枠組みを示している。これまで高次元小標本の理論は出来ていたので、合わせて大規模データベースの挙動が明らかとなる。これは研究目的に即した重要な進展と言える。その他、ベル多項式分布の母数をセル母数に変換した場合、分散母数に依存しない分布族の存在を示した。ベル多項式分布のギブス確率分割への分布収束を示す議論では分散母数について極限をとるので、このような例外を理解出来れば新しい理論展開につながると期待している。
コロナウィルス流行のためグループとしての研究活動は制約を受けたが、流行が弱まった年度末にグループ内研究会を対面で開催した。またオンラインで公開研究集会を開催した。オンラインならではの取り組みとして、英語セッションを設けてイタリアの研究者に講演してもらい、研究代表者等の講演も聞いてもらった。またSpringerから英文研究書籍を出版するなど、研究成果の普及に努めた。

Current Status of Research Progress

2: Research has progressed on the whole more than it was originally planned.

Reason

本計画でベル多項式分布族に関する大標本理論の枠組みを得られるとは予想していなかった。大規模なデータベースについての個票開示リスク評価にまで成果が適用出来ることになり、実績として申し分ない。

Strategy for Future Research Activity

本研究計画も最終年度を迎える。理論的成果は出ているが、論文執筆が追い付いていない。口頭発表による成果普及を先行させるが、論文公刊のための時間を確保したい。

Research Products

• [Int'l Joint Research] rutgers university/university of new hampshire(米国)
• [Int'l Joint Research] Polytechnic University of Catalonia(スペイン)
• [Int'l Joint Research] Rutgers university/The city university of new york(米国)
• [Int'l Joint Research] Rutgers university(米国)
• [Int'l Joint Research] Vassar college/Rutgers university(米国)
• [Int'l Joint Research] Seoul national university(韓国)
• [Journal Article] Sequential modeling, monitoring and forecasting of streaming web traffic data (2022)
  • Author(s): Irie, K., Glynn, C. and Aktekin, T.
  • Journal Title: Annals of Applied Statistics Volume: 26 Pages: 300-325
  • Peer Reviewed / Open Access / Int'l Joint Research
• [Journal Article] On global-local shrinkage for count data (2022)
  • Author(s): Hamura, Y., Irie, K., and Sugasawa, S.
  • Journal Title: Bayesian Analysis Volume: To appear
  • Peer Reviewed / Open Access
• [Journal Article] Score-adjusted methods for estimation of shape parameters in Gamma-Poisson and Beta-Binomial distributions (2022)
  • Author(s): Tamae, H., Irie, K. and Kubokawa, T.
  • Journal Title: Communication in Statistics - Simulation and Computation Volume: To appear
  • Peer Reviewed
• [Journal Article] On admissible estimation of a mean vector when the scale is unknown (2022)
  • Author(s): Maruyama, Y. and Strawderman, W,E.
  • Journal Title: Bernoulli Volume: To appear
  • Peer Reviewed / Int'l Joint Research
• [Journal Article] ノイズを挿入した個票データに対する多重寸法指標を用いた種々のリスク評価方法の比較 (2022)
  • Author(s): 佐井至道
  • Journal Title: 岡山商大論叢 Volume: 58
  • Peer Reviewed / Open Access
• [Journal Article] On the relationship between beta-Bartlett and Uhlig extended processes (2021)
  • Author(s): Pena, V. and Irie, K.
  • Journal Title: Journal of Time Series Analysis Volume: 43 Pages: 147-153
  • Peer Reviewed / Int'l Joint Research
• [Journal Article] Contributed discussion of: Multilevel linear models, Gibbs samplers and multigrid decompositions, by Zanella, G. and Roberts, G. (2021)
  • Author(s): Irie, K. and Sugasawa, S.
  • Journal Title: Bayesian Analysis Volume: 16 Pages: 1378-1380
  • Peer Reviewed / Open Access
• [Journal Article] Admissible estimators of a multivariate normal mean vector when the scale is unknown (2021)
  • Author(s): Maruyama, Y. and Strawderman, W,E.
  • Journal Title: Biometrika Volume: 108 Pages: 997-1003
  • DOI: 10.1093/biomet/asaa102
  • Peer Reviewed / Int'l Joint Research
• [Journal Article] 局所的な分布を用いたリスク評価 －実用化に向けた検討－ (2021)
  • Author(s): 佐井至道
  • Journal Title: 岡山商大論叢 Volume: 56 Pages: 43-76
  • NAID: 120007176717
  • Peer Reviewed
• [Journal Article] A Gaussian sequence approach for proving minimaxity: A Review (2021)
  • Author(s): Maruyama Yuzo、Strawderman William E.
  • Journal Title: Journal of Statistical Planning and Inference Volume: 211 Pages: 256-270
  • DOI: 10.1016/j.jspi.2020.06.007
  • NAID: 120006897876
  • Peer Reviewed / Open Access / Int'l Joint Research
• [Journal Article] A firm foundation for statistical disclosure control (2020)
  • Author(s): Hoshino Nobuaki
  • Journal Title: Japanese Journal of Statistics and Data Science Volume: 3 Issue: 2 Pages: 721-746
  • DOI: 10.1007/s42081-020-00086-9
  • NAID: 210000163233
  • ISSN: 2520-8756, 2520-8764
  • Peer Reviewed
• [Journal Article] Admissible Bayes equivariant estimation of location vectors for spherically symmetric distributions with unknown scale (2020)
  • Author(s): Maruyama Yuzo、Strawderman William E.
  • Journal Title: The Annals of Statistics Volume: 48 Pages: 1052-1071
  • DOI: 10.1214/19-aos1837
  • Peer Reviewed / Open Access / Int'l Joint Research
• [Journal Article] Firm foundation for statistical disclosure control (2020)
  • Author(s): Hoshino Nobuaki
  • Journal Title: Discussion paper, Faculty of economics and management, Kanazawa University Volume: 52 Pages: 1-22
  • NAID: 210000163233
  • Open Access
• [Journal Article] 計算代数のdirect samplerへの応用 (2020)
  • Author(s): 高山信毅，間野修平
  • Journal Title: 数理解析研究所講究録 Volume: 2138 Pages: 64-72
  • NAID: 40022147274
  • Open Access
• [Journal Article] 抽出法と計算代数 (2020)
  • Author(s): 間野修平
  • Journal Title: 統計数理 Volume: 67 Pages: 72-73
  • Open Access
• [Journal Article] 局所的な分布を用いた個票データのリスク評価 (2019)
  • Author(s): 佐井至道
  • Journal Title: 岡山商大論叢 Volume: 55 Pages: 45-70
  • Peer Reviewed
• [Journal Article] Harmonic Bayesian Prediction Under α -Divergence (2019)
  • Author(s): Yuzo Maruyama, Takeru Matsuda, Toshio Ohnishi
  • Journal Title: IEEE Transactions on Information Theory Volume: 65 Pages: 5352-5366
  • DOI: 10.1109/tit.2019.2915245
  • NAID: 120007017124
  • Peer Reviewed / Open Access / Int'l Joint Research
• [Journal Article] The Quasi-Multinomial Synthesizer for Categorical Data (2018)
  • Author(s): Hu Jingchen、Hoshino Nobuaki
  • Journal Title: Privacy in Statistical Databases, PSD 2018, LNCS. Volume: 11126 Pages: 75-91
  • DOI: 10.1007/978-3-319-99771-1_6
  • Peer Reviewed / Int'l Joint Research
• [Journal Article] 個票データのキー変数の型とリスクとの関係 (2018)
  • Author(s): 佐井至道
  • Journal Title: 岡山商大論叢 Volume: 54 Pages: 1-28
• [Journal Article] Posterior consistency of g-prior for variable selection with a growing number of parameters (2018)
  • Author(s): Wang, M. and Maruyama, Y.
  • Journal Title: Journal of Statistical Planning and Inference Volume: 196 Pages: 19-29
  • DOI: 10.1016/j.jspi.2017.10.007
  • Peer Reviewed / Open Access / Int'l Joint Research
• [Presentation] Urn models closed under resizing (2021)
  • Author(s): 星野伸明
  • Organizer: 2021年度統計関連学会連合大会
• [Presentation] Bayesian dynamic fused lasso (2021)
  • Author(s): Kaoru Irie
  • Organizer: Bayesian Inference in Stochastic Processes 12 (BISP12)
  • Int'l Joint Research / Invited
• [Presentation] Robust regression with log-regularly varying error distributions (2021)
  • Author(s): Kaoru Irie (Yasuyuki Hamura and Shonosuke Sugasawa)
  • Organizer: Eastern Asia Chapter of the International Society for Bayesian Analysis (EAC-ISBA 2021)
  • Int'l Joint Research / Invited
• [Presentation] On the conjugate multivariate stochastic volatility processes (2021)
  • Author(s): Kaoru Irie (Victor Pena)
  • Organizer: The 14th International Conference of the ERCIM WG on Computational and Methodological Statistics (CMStatistics 2021)
  • Int'l Joint Research / Invited
• [Presentation] 単調回帰のための事前分布 (2021)
  • Author(s): 入江薫（岡野遼、羽村靖之、菅澤翔之助）
  • Organizer: 2021年度統計関連学会連合大会
• [Presentation] 漸近不偏推定量の幾何的構成 (2021)
  • Author(s): 廣瀬雅代, 間野修平
  • Organizer: 2021年度統計関連学会連合大会
• [Presentation] A Bayesian construction of asymptotic unbiased estimators (2021)
  • Author(s): Hirose, M. and Mano, S.
  • Organizer: 2021 World Meeting of International Society for Bayesian Analysis
  • Int'l Joint Research
• [Presentation] Admissible estimators of a multivariate normal mean vector when the scale is unknown (2021)
  • Author(s): Yuzo Maruyama
  • Organizer: Eastern Asia Chapter of the International Society for Bayesian Analysis (EAC-ISBA 2021)
  • Int'l Joint Research / Invited
• [Presentation] 平均ベクトルの推定における分散未知のもとでの許容的でミニマクスな推定量 (2021)
  • Author(s): 丸山祐造
  • Organizer: 2021年度統計関連学会連合大会
• [Presentation] ノイズが挿入された個票データに対する 多重寸法指標を用いた いくつかのリスク評価方法の比較 (2021)
  • Author(s): 佐井至道
  • Organizer: 2021年度統計関連学会連合大会
• [Presentation] Robust regression with log-regularly varying error distributions (2021)
  • Author(s): Kaoru Irie, Yasuyuki Hamura, Shonosuke Sugasawa
  • Organizer: 日本統計学会春季大会
  • Invited
• [Presentation] 差分プライバシーの母数の決め方 (2020)
  • Author(s): 星野伸明
  • Organizer: 2020年度統計関連学会連合大会
• [Presentation] 局所的な分布を用いたリスク評価の実用化に向けた検討 (2020)
  • Author(s): 佐井至道
  • Organizer: 2020年度統計関連学会連合大会
• [Presentation] On the conjugate multivariate stochastic volatility processes (2020)
  • Author(s): Kaoru Irie, Victor Pena
  • Organizer: Webinar of Bayesian Econometrics 2020
  • Int'l Joint Research / Invited
• [Presentation] 差分プライベートなサンプリング (2019)
  • Author(s): 星野伸明
  • Organizer: 2019年度統計関連学会連合大会
• [Presentation] 差分プライベートなミクロデータ (2019)
  • Author(s): 星野伸明
  • Organizer: Computer Security Symposium 2019
  • Invited
• [Presentation] 局所的な分布を用いた個票データのリスク評価 (2019)
  • Author(s): 佐井至道
  • Organizer: 2019年度統計関連学会連合大会
• [Presentation] Ensemble minimaxity of James-Stein estimators (2019)
  • Author(s): Yuzo Maruyama
  • Organizer: Symposium in Memory of Charles Stein
  • Int'l Joint Research / Invited
• [Presentation] Ensemble minimaxity of James-Stein estimators (2019)
  • Author(s): Yuzo Maruyama
  • Organizer: "New and Evolving Roles of Shrinkage in Large-Scale Prediction and Inference" workshop
  • Int'l Joint Research / Invited
• [Presentation] Samplers from toric models with computational algebra (2019)
  • Author(s): Shuhei Mano
  • Organizer: Algebraic Statistics Workshop, Research Institute for Computational and Applied Mathematics, Linz, Austria
  • Int'l Joint Research / Invited
• [Presentation] Direct sampler with computational algebra for toric models (2019)
  • Author(s): Shuhei Mano
  • Organizer: The 28th South Taiwan Statistics Conference, Taichun, Taiwan
  • Int'l Joint Research / Invited
• [Presentation] 匿名のデータとは何か (2018)
  • Author(s): 星野伸明
  • Organizer: 日本計算機統計学会第32回シンポジウム
  • Invited
• [Presentation] 統計的推測精度の管理 (2018)
  • Author(s): 星野伸明
  • Organizer: Computer Security Symposium 2018
  • Invited
• [Presentation] Rejection Sampling from the Quasi-Multinomial Distribution (2018)
  • Author(s): 星野伸明
  • Organizer: 2018年度統計関連学会連合大会
• [Presentation] The Quasi-Multinomial Synthesizer for Categorical Data (2018)
  • Author(s): Hu, J. and Hoshino, N.
  • Organizer: Privacy in Statistical Databases (PSD) 2018
  • Int'l Joint Research
• [Presentation] リスクの高いデータの秘匿について (2018)
  • Author(s): 佐井至道
  • Organizer: 2018年度統計関連学会連合大会
• [Presentation] 非攪乱的および攪乱的な秘匿方法を施した個票データのリスクの比較 (2018)
  • Author(s): 佐井至道
  • Organizer: Computer Security Symposium 2018
  • Invited
• [Presentation] Harmonic Bayesian prediction under alpha-divergence (2018)
  • Author(s): Yuzo Maruyama
  • Organizer: The 3rd Eastern Asia meeting on Bayesian Statistics
  • Int'l Joint Research / Invited
• [Presentation] Samplers for non-exchangeable prior processes and hypergeometric systems (2018)
  • Author(s): Shuhei Mano
  • Organizer: The 5th IMS-APRM meeting
  • Int'l Joint Research / Invited
• [Presentation] Sampling from non-exchangeable priors (2018)
  • Author(s): Shuhei Mano
  • Organizer: EAC-ISBA 2018 Conference
  • Int'l Joint Research
• [Presentation] Direct Sequential Sampler for A-hypergeometric Distributions (2018)
  • Author(s): Shuhei Mano and Nobuki Takayama
  • Organizer: IASC-ARS and CASC joint conference
  • Int'l Joint Research / Invited
• [Presentation] A direct sampler from A-hypergeometric distributions (2018)
  • Author(s): Shuhei Mano and Nobuki Takayama
  • Organizer: 2018年度統計関連学会連合大会
• [Presentation] 分割，超幾何系，Dirichlet過程と統計的推測 (2018)
  • Author(s): 間野修平
  • Organizer: 日本数学会2018年度秋季総合分科会
  • Invited
• [Book] Pioneering Works on Extreme Value Theory (2021)
  • Author(s): Nobuaki Hoshino, Shuhei Mano, and Takaaki Shimura (Eds.)
  • Total Pages: 143
  • Publisher: Springer, Singapore
  • ISBN: 9789811607677
• [Book] Pioneering Works on Distribution Theory (2021)
  • Author(s): Hoshino, N., Mano, S., and Shimura, T. (Eds.)
  • Total Pages: 128
  • Publisher: Springer
  • ISBN: 981159662X
• [Book] 日本統計学会公式認定 統計検定準1級対応 統計学実践ワークブック (2020)
  • Author(s): 竹村彰通他編著，青木敏，佐井至道，その他18名
  • Total Pages: 352
  • Publisher: 学術図書出版社
  • ISBN: 9784780608526
• [Book] Partitions, Hypergeometric Systems, and Dirichlet Processes in Statistics (2018)
  • Author(s): Shuhei Mano
  • Total Pages: 135
  • Publisher: Springer
  • ISBN: 9784431558866
• [Funded Workshop] 大規模データの公開におけるプライバシー保護の理論と応用 (2021)
• [Funded Workshop] Pioneering Workshop on Extreme Value and Distribution Theories (2019)