Abstract
Contingency tables have provided a fertile ground for the growth of algebraic statistics. In this paper we briefly outline some features of this work and point to open research problems. We focus on the problem of maximum likelihood estimation for log-linear models and a related problem of disclosure limitation to protect the confidentiality of individual responses. Risk of disclosure has often been measured either formally or informally in terms of information contained in marginal tables linked to a log-linear model and has focused on the disclosure potential of small cell counts, especially those equal to 1 or 2. One way to assess the risk is to compute bounds for cell entries given a set of released marginals. Both of these methodologies become complicated for large sparse tables. This paper revisits the problem of computing bounds for cell entries and picks up on a theme first suggested in Fienberg [21] that there is an intimate link between the ideas on bounds and the existence of maximum likelihood estimates, and shows how these ideas can be made rigorous through the underlying mathematics of the same geometric/algebraic framework. We illustrate the linkages through a series of examples. We also discuss the more complex problem of releasing marginal and conditional information. We illustrate the statistical features of the methodology on two examples and then conclude with a series of open problems.
AMS(MOS) subject classifications. 13P10, 62805, 62H17, 62P25.
Supported in part by NSF grants E1A9876619 and 11S0131884 to the National Institute of Statistical Sciences, and NSF Grant DMS-0631589 and Army contract DAAD19-02-1-3-0389 to Carnegie Mellon University.
Supported in part by NSF Grant DMS-0631589 and a grant from the Pennsylvania Department of Health through the Commonwealth Universal Research Enhancement Program to Carnegie Mellon University.
Supported in part by NSF grants EIA9876619 and 11S0131884 to the National Institute of Statistical Sciences and SES-0532407 to Pennsylvania State University.
Supported by Army contract DAAD19-02-1-3-0389 to Carnegie Mellon University.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Y.M.M. BISHOP, S.E. FIENBERG, AND P.W. HOLLAND (1975). Discrete Multivariate Analysis: Theory and Practice, MIT Press, Cambridge, MA. Reprinted (2007), Springer-Verlag, New York.
L. BUZZIGOLI AND A. GIUSTI (1999). An algorithm to calculate the lower and upper bounds of the elements of an array given its marginals, in Proceedings of the Conference on Statistical Data Protection. Luxemburg: Eurostat, pp. 131-147.
E. CARLINI AND F. RAPALLO (2005). ThegeometryofstaUstical models for two-way contingency tables with fixed odds ratios, Rendiconti dell'Istituto di Matematica dell'Universita di Trieste, 37:71~84.
S.K. CHRISTIANSEN AND H. GIESE (19991). Genetic analysis of obligate barley powdery mildew fungus based on RFPL and virulence loci, Theoretical and Applied Genetics, 79:705-712.
Y. CHEN, I.R. DINWOODIE, AND S. SULLIVANT (2006). Sequential importance sampling for multiway tables, Annals of Statistics, 34:523-545.
L.R. Cox (2002). Bounds on entries in 3-dimensional contingency tables subject to given marginal totals, In J. Domingo-Ferrer (Ed.), Inference Oontrol in Statistical Databases, Springer-Verlag LNCS 2316, pp. 21-33.
L.R. Cox (2003). On properties of multi-dimensional statistical tables, Journal of Statistical Planning and Inference, 117:251-273.
J.A. DE LOERA, R. HEMMECKE, J. TAUZER, AND R. YOSHIDA (2004). Effective lattice point counting in rational convex polytopes, Journal of Symbolic Computation, 38:1273-1302.
J.A. DE LOERA AND S. ONN (2006). Markov bases of 3-way tables are erbitrerily' complicated, Journal of Symbolic Computation, 41:173-181.
[101 P. DIACONIS AND B. STURMFELS (1998). Algebraic algorUbms for sampling from conditional distribution, Annals of Statistics, 26:363-397.
[11J A. DOBRA (2002). Statistical Tools for Disclosure Limitation in Multi~way Contingency Tables. Ph.D. Dissertation, Department of Statistics, Carnegie Mellon University.
[12J A. DOBRA (2003). Markov bases for decomposable graphical models, Bernoulli, 9(6):1-16.
A. DOBRA AND S.E. FIENBERG (2000). Bounds for cell entries iti contingency tables given marginal totals and decomposable graphs, Proceedings of the National Academy of Sciences, 97:11885-11892.
A. DOBRAAND S.E. FIENBERG (2001). "Bounds for cell entries in contingency tables induced by fixed marginal totals with applications to disclosure limitation," Statistical Journal of the United Nations ECE, 18:363-37l.
A. DOBRAAND S.E. FIENBERG (2003). Bounding entries in multi-way contingency tables given a set of marginal totals, in Y. Haitovsky, H.R. Lerche, and Y. Ritov, eds., Foundations of Statistical Inference: Proceedings of the Shoresh Conference 2000, Physica-Verlag, pp. 3-16.
A. DOBRA AND S.E. FIENBERG (2008). The generalized shuttle algorithm, in P. Gibilisco, Eva Riccomagno, Maria-Piera Rogantin (eds.) Algebraic and Geometric Methods in Probability and Statistics, Cambridge University Press, to appear.
[17) A. DOBRA, S.E. FIENBERG, AND M. TROTTINI (2003). Assessing the risk of disclosure of confidential categorical data, in J. Bernardo et al., eds., Bayesian StatisUcs 7, Oxford University Press, pp. 125-144.
[18J P. DOYLE, J. LANE, J. THEEUWES, AND L. ZAYATZ (eds.) (2001). Confidentiality, Disclosure and Data Access: Theory and Practical Applications for StatistJ'cal Agencies. Elsevier.
[191 D. EDWARDS (1992). Linkage analysis using log-lineet models, Computational Statistics and Data Analysis, 10:281-290.
N. ERIKSSON, S.E. FIENBERG, A. RJNALDO, AND S. SULLIVANT (2006). Polyhedral conditions for the non-existence of the MLE for hierarchical log-linear models, Journal of Symbolic Computation, 41:222-233.
S.E. FIENBERG (1999). Frechet and Bonferroni bounds for multi-way tables of counts With applications to disclosure limitation, In Statistical Data Protection, Proceedings of the Conference, Lisbon, Eurostat, pp. 115-13l.
S.E. FIENBERG, U.E. MAKOV, M.M. MEYER, AND R.J. STEELE (2001). "Computing the exact distribution for a multi-way contingency table conditional on its marginal totals," in A.K.M.E. Saleh, ed., Data Analysis from Statistica Foundations: A Festschrift in Honor of the 75th Birthday of D. A. S. Fraser, Nova Science Publishers, Huntington, NY, pp. 145-165.
S.E. FIENBERG AND A. RINALDO (2006). Computing maximum likelihood estimates in log-linear models, Technical Report 835, Department of Statistics, Carnegie Mellon University.
S.E. FIENBERG AND A. RINALDO (2007). Three centuries of categorical data analysis: log-linear models and maximum likelihood estimation, Journal of Statistical Planning and Inference, 137:3430-3445.
S.E. FIENBERG AND A.B. SLAVKOVIC (2004a). Making the release of confidential data from multi-way tables count, Chance, 17(3):5-10.
S.E FIENBERG AND A.S. SLAVKOVIC (2005). Preserving the confidentiality of categorical databases when releasing information for association rules, Data Mining and Knowledge Discovery, 11:155-180.
L. GARCIA, M. STILLMAN, AND B. STURMFELS (2005). Algebraic geometry for Bayesian networks, Journal of Symbolic Computation, 39:331-355.
E. GAWRILOW AND M. JOSWIG (2005). Geometric reasoning with polymake, Manuscript available at arXiv:math. CO/0507273.
D. GEIGER, C. MEEK, AND B. STURMFELS (2006). On the toric algebra ofgraphical models, Annals of Statistics, 34:1463-1492.
S.J. HABERMAN (1974). The Analysis of Frequency Data, University of Chicago Press, Chicago, Illinois.
S. HO~TEN AND B. STURMFELS (2006). Computing the integer programming gap, Combinatorica, 27:367-382.
S.L. LAURJTZEN (1996). Graphical Models, Oxford University Press, New York. [33] R.B. NELSEN (2006). An Introduction to Copulas. Springer-Verlag, New York.
A. RINALDO (2005). Maximum Likelihood Estimation for Log-linear Models. Ph.D. Dissertation, Department of Statistics, Carnegie Mellon University.
A. RINALDO (2006). On maximum likelihood estimation for log-linear models, submitted for publication.
F. SANTOS AND B. STURMFELS (2003). Higher Lawrence configurations, J. Cornbin. Theory Ser. A, 103:151-164.
A.B. SLAVKOVIC (2004). Statistical Disclosure Limitation Beyond the Margins: Characterization of Joint Distributions for Contingency Tables. Ph.D. Dissertation, Department of Statistics, Carnegie Mellon University.
A.B. SLAVKOVIC AND B. SMUCKER (2007). Calculating Cell Bounds in Contingency Tables Based on Conditional Frequencies, Technical Report, Department of Statistics, Penn State University.
A.B. SLAVKOVIC AND FIENBERG, S. E. (2004). Bounds for Cell Entries in Two-way Tables Given Conditional Relative Frequencies, In Domingo-Ferrer, J. and Terra, V. (eds.), Privacy in Statistical Databases, Lecture Notes in Computer Science No. 3050, pp. 30-43. New York: Springer-Verlag.
A.B. SLAVKOVIC AND S.E. FIENBERG (2008). The algebraic geometry of 2 x 2 contingency tables, forthcoming.
B. STURMFELS (1995). Grebner Bases and Convex Polytope, American Mathematical Society, University Lecture Series, 8.
S. SULL1VANT (2006). Compressed polytopes and statistical disclosure limitation, Tohoku Mathematical Journal, 58(3):433-445.
S. SULLIVANT (2005). Small contingency tables with large gaps, SIAM Journal of Discrete Mathematics, 18(4):787-793.
G.M. ZIEGLER (1998). Lectures on Polytopes, Springer-Verlag, New York.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2009 Springer Science+Business Media, LLC
About this chapter
Cite this chapter
Dobra, A., Fienberg, S.E., Rinaldo, A., Slavkovic, A., Zhou, Y. (2009). Algebraic Statistics and Contingency Table Problems: Log-Linear Models, Likelihood Estimation, and Disclosure Limitation. In: Putinar, M., Sullivant, S. (eds) Emerging Applications of Algebraic Geometry. The IMA Volumes in Mathematics and its Applications, vol 149. Springer, New York, NY. https://doi.org/10.1007/978-0-387-09686-5_3
Download citation
DOI: https://doi.org/10.1007/978-0-387-09686-5_3
Published:
Publisher Name: Springer, New York, NY
Print ISBN: 978-0-387-09685-8
Online ISBN: 978-0-387-09686-5
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)