IDEAS home Printed from https://ideas.repec.org/p/arx/papers/1907.01954.html
   My bibliography  Save this paper

An Econometric Perspective on Algorithmic Subsampling

Author

Listed:
  • Sokbae Lee
  • Serena Ng

Abstract

Datasets that are terabytes in size are increasingly common, but computer bottlenecks often frustrate a complete analysis of the data. While more data are better than less, diminishing returns suggest that we may not need terabytes of data to estimate a parameter or test a hypothesis. But which rows of data should we analyze, and might an arbitrary subset of rows preserve the features of the original data? This paper reviews a line of work that is grounded in theoretical computer science and numerical linear algebra, and which finds that an algorithmically desirable sketch, which is a randomly chosen subset of the data, must preserve the eigenstructure of the data, a property known as a subspace embedding. Building on this work, we study how prediction and inference can be affected by data sketching within a linear regression setup. We show that the sketching error is small compared to the sample size effect which a researcher can control. As a sketch size that is algorithmically optimal may not be suitable for prediction and inference, we use statistical arguments to provide 'inference conscious' guides to the sketch size. When appropriately implemented, an estimator that pools over different sketches can be nearly as efficient as the infeasible one using the full sample.

Suggested Citation

  • Sokbae Lee & Serena Ng, 2019. "An Econometric Perspective on Algorithmic Subsampling," Papers 1907.01954, arXiv.org, revised Apr 2020.
  • Handle: RePEc:arx:papers:1907.01954
    as

    Download full text from publisher

    File URL: http://arxiv.org/pdf/1907.01954
    File Function: Latest version
    Download Restriction: no
    ---><---

    Other versions of this item:

    References listed on IDEAS

    as
    1. Boivin, Jean & Ng, Serena, 2006. "Are more data always better for factor analysis?," Journal of Econometrics, Elsevier, vol. 132(1), pages 169-194, May.
    2. Bai, Z. D. & Silverstein, Jack W. & Yin, Y. Q., 1988. "A note on the largest eigenvalue of a large dimensional sample covariance matrix," Journal of Multivariate Analysis, Elsevier, vol. 26(2), pages 166-168, August.
    3. Cramer, J. S., 1987. "Mean and variance of R2 in small and moderate samples," Journal of Econometrics, Elsevier, vol. 35(2-3), pages 253-266, July.
    4. HaiYing Wang & Min Yang & John Stufken, 2019. "Information-Based Optimal Subdata Selection for Big Data Linear Regression," Journal of the American Statistical Association, Taylor & Francis Journals, vol. 114(525), pages 393-405, January.
    5. Wallace, T D, 1972. "Weaker Criteria and Tests for Linear Restrictions in Regression," Econometrica, Econometric Society, vol. 40(4), pages 689-698, July.
    6. HaiYing Wang & Rong Zhu & Ping Ma, 2018. "Optimal Subsampling for Large Sample Logistic Regression," Journal of the American Statistical Association, Taylor & Francis Journals, vol. 113(522), pages 829-844, April.
    7. Serena Ng, 2017. "Opportunities and Challenges: Lessons from Analyzing Terabytes of Scanner Data," NBER Working Papers 23673, National Bureau of Economic Research, Inc.
    8. I. T. Jolliffe, 1972. "Discarding Variables in a Principal Component Analysis. I: Artificial Data," Journal of the Royal Statistical Society Series C, Royal Statistical Society, vol. 21(2), pages 160-173, June.
    9. Christmann, Andreas & Steinwart, Ingo & Hubert, Mia, 2007. "Robust learning from bites for data mining," Computational Statistics & Data Analysis, Elsevier, vol. 52(1), pages 347-361, September.
    Full references (including those not matched with items on IDEAS)

    Citations

    Citations are extracted by the CitEc Project, subscribe to its RSS feed for this item.
    as


    Cited by:

    1. Tao Zou & Xian Li & Xuan Liang & Hansheng Wang, 2021. "On the Subbagging Estimation for Massive Data," Papers 2103.00631, arXiv.org.
    2. Jun Yu & HaiYing Wang, 2022. "Subdata selection algorithm for linear model discrimination," Statistical Papers, Springer, vol. 63(6), pages 1883-1906, December.
    3. Sokbae Lee & Yuan Liao & Myung Hwan Seo & Youngki Shin, 2022. "Fast Inference for Quantile Regression with Tens of Millions of Observations," Papers 2209.14502, arXiv.org, revised Oct 2023.
    4. Sokbae Lee & Serena Ng, 2020. "Least Squares Estimation Using Sketched Data with Heteroskedastic Errors," Papers 2007.07781, arXiv.org, revised Jun 2022.
    5. O’Connell, Martin & Smith, Howard & Thomassen, Øyvind, 2023. "A two sample size estimator for large data sets," Discussion Papers 2023/1, Norwegian School of Economics, Department of Business and Management Science.

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Jun Yu & HaiYing Wang, 2022. "Subdata selection algorithm for linear model discrimination," Statistical Papers, Springer, vol. 63(6), pages 1883-1906, December.
    2. Tao Zou & Xian Li & Xuan Liang & Hansheng Wang, 2021. "On the Subbagging Estimation for Massive Data," Papers 2103.00631, arXiv.org.
    3. Onatski, A., 2018. "Asymptotics of the principal components estimator of large factor models with weak factors and i.i.d. Gaussian noise," Cambridge Working Papers in Economics 1808, Faculty of Economics, University of Cambridge.
    4. Laurent Ferrara & Anna Simoni, 2023. "When are Google Data Useful to Nowcast GDP? An Approach via Preselection and Shrinkage," Journal of Business & Economic Statistics, Taylor & Francis Journals, vol. 41(4), pages 1188-1202, October.
    5. Feifei Wang & Danyang Huang & Tianchen Gao & Shuyuan Wu & Hansheng Wang, 2022. "Sequential one‐step estimator by sub‐sampling for customer churn analysis with massive data sets," Journal of the Royal Statistical Society Series C, Royal Statistical Society, vol. 71(5), pages 1753-1786, November.
    6. Su, Miaomiao & Wang, Ruoyu & Wang, Qihua, 2022. "A two-stage optimal subsampling estimation for missing data problems with large-scale data," Computational Statistics & Data Analysis, Elsevier, vol. 173(C).
    7. Jun Yu & Jiaqi Liu & HaiYing Wang, 2023. "Information-based optimal subdata selection for non-linear models," Statistical Papers, Springer, vol. 64(4), pages 1069-1093, August.
    8. Duarte, Belmiro P.M. & Atkinson, Anthony C. & Oliveira, Nuno M.C., 2024. "Using hierarchical information-theoretic criteria to optimize subsampling of extensive datasets," LSE Research Online Documents on Economics 121641, London School of Economics and Political Science, LSE Library.
    9. Tianzhen Wang & Haixiang Zhang, 2022. "Optimal subsampling for multiplicative regression with massive data," Statistica Neerlandica, Netherlands Society for Statistics and Operations Research, vol. 76(4), pages 418-449, November.
    10. J. Lars Kirkby & Dang H. Nguyen & Duy Nguyen & Nhu N. Nguyen, 2022. "Inversion-free subsampling Newton’s method for large sample logistic regression," Statistical Papers, Springer, vol. 63(3), pages 943-963, June.
    11. Ziyang Wang & HaiYing Wang & Nalini Ravishanker, 2023. "Subsampling in Longitudinal Models," Methodology and Computing in Applied Probability, Springer, vol. 25(1), pages 1-29, March.
    12. Amalan Mahendran & Helen Thompson & James M. McGree, 2023. "A model robust subsampling approach for Generalised Linear Models in big data settings," Statistical Papers, Springer, vol. 64(4), pages 1137-1157, August.
    13. Zhang, Haixiang & Wang, HaiYing, 2021. "Distributed subdata selection for big data via sampling-based approach," Computational Statistics & Data Analysis, Elsevier, vol. 153(C).
    14. Hector, Emily C. & Luo, Lan & Song, Peter X.-K., 2023. "Parallel-and-stream accelerator for computationally fast supervised learning," Computational Statistics & Data Analysis, Elsevier, vol. 177(C).
    15. Lee, JooChul & Wang, HaiYing & Schifano, Elizabeth D., 2020. "Online updating method to correct for measurement error in big data streams," Computational Statistics & Data Analysis, Elsevier, vol. 149(C).
    16. Serena Ng & Susannah Scanlan, 2023. "Constructing High Frequency Economic Indicators by Imputation," Papers 2303.01863, arXiv.org, revised Oct 2023.
    17. Xiaohui Yuan & Yong Li & Xiaogang Dong & Tianqing Liu, 2022. "Optimal subsampling for composite quantile regression in big data," Statistical Papers, Springer, vol. 63(5), pages 1649-1676, October.
    18. Lulu Zuo & Haixiang Zhang & HaiYing Wang & Liuquan Sun, 2021. "Optimal subsample selection for massive logistic regression with distributed data," Computational Statistics, Springer, vol. 36(4), pages 2535-2562, December.
    19. Martin Schneider & Martin Spitzer, 2004. "Forecasting Austrian GDP using the generalized dynamic factor model," Working Papers 89, Oesterreichische Nationalbank (Austrian Central Bank).
    20. Oxana Babecka Kucharcukova & Jan Bruha, 2016. "Nowcasting the Czech Trade Balance," Working Papers 2016/11, Czech National Bank.

    More about this item

    NEP fields

    This paper has been announced in the following NEP Reports:

    Statistics

    Access and download statistics

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:arx:papers:1907.01954. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: arXiv administrators (email available below). General contact details of provider: http://arxiv.org/ .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.