zbMATH — the first resource for mathematics

Sparse estimation of multivariate Poisson log-normal models from count data. (English) Zbl 07260731
Summary: Modeling data with multivariate count responses is a challenging problem because of the discrete nature of the responses. Existing methods for univariate count response cannot be easily extended to the multivariate case since the dependence among multiple responses needs to be properly accommodated. In this paper, we propose a multivariate Poisson log-normal regression model for multivariate count responses by using latent variables. By simultaneously estimating the regression coefficients and inverse covariance matrix over the latent variables with an efficient Monte Carlo EM algorithm, the proposed model takes advantage of the association among multiple count responses to improve the model prediction accuracy. Simulation studies and applications to real-world data are conducted to systematically evaluate the performance of the proposed method in comparison with conventional methods.

62-XX Statistics
68-XX Computer science
glmnet; glasso; EMBERS
Full Text: DOI arXiv
[1] Argyriou, A., Evgeniou, T., and Pontil, M., Multi‐task feature learning, NIPS, 2007, pp. 41-48.
[2] Argyriou, A., Micchelli, C. A., Pontil, M., and Ying, Y., A spectral regularization framework for multi‐task structure learning, NIPS, 2007.
[3] Blei, D. M., Kucubelbir, A., and McAuliffe, J. D., Variational inference: A review for statisticians, 2016, available at https://arxiv.org/abs/1601.00670.
[4] Chen, J. and Chen, Z., Extended Bayesian information criteria for model selection with large model spaces, Biometrika95( 3) ( 2008), 759– 771. MR2443189 · Zbl 1437.62415
[5] Chen, J., Zhou, J., and Ye, J., Integrating low‐rank and group‐sparse structures for robust multi‐task learning, KDD ‘11, 2011, pp. 42-50.
[6] Chib, S., Greenberg, E., and Winkelmann, R., Posterior simulation and bayes factors in panel count data models, J. Econometrics86( 1) ( 1998), 33– 54. · Zbl 1041.62524
[7] El‐Basyouny, K. and Sayed, T., Collision prediction models using multivariate Poisson‐lognormal regression, Accid. Anal. Prev.41( 4) ( 2009), 820– 828.
[8] Foygel, R. and Drton, M., Extended Bayesian information criteria for Gaussian graphical models, NIPS, 2010, pp. 604-612.
[9] Friedman, J. et al., Pathwise coordinate optimization, Ann. Appl. Stat.1( 2) ( 2007), 302– 332. MR2415737 · Zbl 1378.90064
[10] Friedman, J., Hastie, T., and Tibshirani, R., Sparse inverse covariance estimation with the graphical lasso, Biostatistics9( 3) ( 2008), 432– 441. · Zbl 1143.62076
[11] Friedman, J., Hastie, T., and Tibshirani, R., The elements of statistical learning, Springer series in statistics, Vol 2, Springer-Verlag, New York, 2009. MR2722294 · Zbl 1273.62005
[12] Friedman, J., Hastie, T., Simon, N., and Tibshirani, R., Lasso and elastic‐net regularized generalized linear models, glmnet R package, 2014.
[13] Gong, P., Ye, J., and Shui Zhang, C., Multi‐stage multi‐task feature learning, NIPS, 2012.
[14] Gong, P., Ye, J., and Zhang, C., Robust multi‐task feature learning, KDD ‘12, 2012, pp. 895-903.
[15] Gong, P., Zhou, J., Fan, W., and Ye, J., Efficient multi‐task feature learning with calibration, KDD ‘14, 2014, pp. 761-770.
[16] Hadiji, F. et al., Poisson dependency networks: Gradient boosted models for multivariate count data, Mach. Learn.100( 2) ( 2015), 477– 507. MR3383979 · Zbl 1388.62218
[17] Higham, N. J., Computing the nearest correlation matrix – A problem from finance, J. Numer. Anal.22( 3) ( 2002), 329– 343. · Zbl 1006.65036
[18] Jalali, A., Sanghavi, S., Ruan, C., and Ravikumar, P. K., A dirty model for multi‐task learning, NIPS, 2010, pp. 964-972.
[19] Karlis, D., An EM algorithm for multivariate Poisson distribution and related models, J. Appl. Stat.30( 1) ( 2003), 63– 77. MR1957361 · Zbl 1121.62408
[20] Karlis, D. and Meligkotsidou, L., Multivariate Poisson regression with covariance structure, Stat. Comput.15( 4) ( 2005), 255– 265. MR2205389
[21] Kumar, A. and Daumé, H., Learning task grouping and overlap in multi‐task learning, ICML ‘12, 2012.
[22] Levina, E. et al., Sparse estimation of large covariance matrices via a nested lasso penalty, Ann. Appl. Stat.2( 1) ( 2008), 245– 263. MR2415602 · Zbl 1137.62338
[23] Liu, H., Wang, L., and Zhao, T., Multivariate regression with calibration, NIPS, 2014, pp. 127-135.
[24] Lozano, A. C., Jiang, H., and Deng, X., Robust sparse estimation of multiresponse regression and inverse covariance matrix via the l2 distance, KDD ‘13, 2013, pp. 293-301.
[25] Ma, J., Kockelman, K. M., and Damien, P., A multivariate Poisson‐lognormal regression model for prediction of crash counts by severity, using Baysian methods, Accid. Anal. Prev.40 ( 2008), 964– 975.
[26] Ramakrishnan, N., Butler, P., Muthiah, S., Self, N., Khandpur, R., Saraf, P., Wang, W., Cadena, J., Vullikanti, A., Korkmaz, G., Kuhlman, C., Marathe, A., Zhao, L., Hua, T., Chen, F., Lu, C. T., Huang, B., Srinivasan, A., Trinh, K., Getoor, L., Katz, G., Doyle, A., Ackermann, C., Zavorin, I., Ford, J., Summers, K., Fayed, Y., Arredondo, J., Gupta, D., and Mares, D., ‘Beating the news’ with embers: Forecasting civil unrest using open source indicators, KDD ‘14, 2014, pp. 1799-1808.
[27] Rothman, A. J., Levina, E., and Zhu, J., Sparse multivariate regression with covariance estimation, J. Comput. Graph. Statist.19( 4) ( 2010), 947– 962. MR2791263
[28] Simon, N. et al., A sparse‐group lasso, J. Comput. Graph. Statist.22( 2) ( 2013), 231– 245. MR3173712
[29] Tibshirani, R., Regression shrinkage and selection via the lasso, J. R. Stat. Soc. Ser. B58 ( 1994), 267– 288. MR2815776 · Zbl 0850.62538
[30] Wang, H., Kalwani, M. U., and Akçura, T., A Bayesian multivariate Poisson regression model of cross‐category store brand purchasing behavior, J. Retailing Consumer Serv.14( 6) ( 2007), 369– 382.
[31] Wang, W., Liang, Y., and Xing, E. P., Block regularized lasso for multivariate multi‐response linear regression, AISTATS, 2013, pp. 608-617.
[32] Wang, Z., Chakraborty, P., Mekaru, S. R., Brownstein, J. S., Ye, J., and Ramakrishnan, N., Dynamic Poisson autoregression for influenza‐like‐illness case count prediction, KDD ‘15, 2015, pp. 1285-1294.
[33] WHO FluNet, 2015, available at http://www.who.int/influenza/gisrs_laboratory/flunet/en/.
[34] Wytock, M. and Kolter, J. Z., Sparse Gaussian conditional random fields: Algorithms, theory, and application to energy forecasting, ICML ‘13, 2013, pp. 1265-1273.
[35] Yang, E., Ravikumar, P. K., Allen, G. I., and Liu, Z., On Poisson graphical models, NIPS ‘13, 2013.
[36] Yu, S., Tresp, V., and Yu, K., Robust multi‐task learning with t‐processes, ICML ‘07, 2007.
[37] Zhao, L., Sun, Q., Ye, J., Chen, F., Lu, C.‐T., and Ramakrishnan, N., Multi‐task learning for spatio‐temporal event forecasting, KDD ‘15, 2015, pp. 1503-1512.
[38] Zoh, R. S. et al., PCAN: Probabilistic correlation analysis of two non‐normal data sets, Biometrics 72( 4) ( 2016), 1358– 1368. MR3591620 · Zbl 1390.62325
[39] Zou, H. and Hastie, T., Regularization and variable selection via the elastic net, J. R. Stat. Soc. Ser. B (Stat. Methodol.)67( 2) ( 2005), 301– 320. MR2137327 · Zbl 1069.62054
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.