Skip to main content
SpringerLink
Log in

Forecasting skewed biased stochastic ozone days: analyses, solutions and beyond

  • Regular Paper
  • Published:
Knowledge and Information Systems Aims and scope Submit manuscript

Abstract

Much work on skewed, stochastic, high dimensional, and biased datasets usually implicitly solve each problem separately. Recently, we have been approached by Texas Commission on Environmental Quality (TCEQ) to help them build highly accurate ozone level alarm forecasting models for the Houston area, where these technical difficulties come together in one single problem. Key characteristics of this problem that is challenging and interesting include: (1) the dataset is sparse (72 features, and 2 or 5% positives depending on the criteria of “ozone days”), (2) evolving over time from year to year, (3) limited in collected data size (7  years or around 2,500 data entries), (4) contains a large number of irrelevant features, (5) is biased in terms of “sample selection bias”, and (6) the true model is stochastic as a function of measurable factors. Besides solving a difficult application problem, this dataset offers a unique opportunity to explore new and existing data mining techniques, and to provide experience, guidance and solution for similar problems. Our main technical focus addresses on how to estimate reliable probability given both sample selection bias and a large number of irrelevant features, and how to choose the most reliable decision threshold to predict the unknown future with different distribution. On the application side, the prediction accuracy of our chosen approach (bagging probabilistic decision trees and random decision trees) is 20% higher in recall (correctly detects 1–3 more ozone days, depending on the year) and 10% higher in precision (15–30 fewer false alarm days per year) than state-of-the-art methods used by air quality control scientists, and these results are significant for TCEQ. On the technical side of data mining, extensive empirical results demonstrate that, at least for this problem, and probably other problems with similar characteristics, these two straight-forward non-parametric methods can provide significantly more accurate and reliable solutions than a number of sophisticated and well-known algorithms, such as SVM and AdaBoost among many others.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  1. Chang CC, Lin CJ (2001) LIBSVM: a library for support vector machines. Software available at http://www.csie.ntu.edu.tw/cjlin/libsvm

  2. Cortes C and Vapnik V (1995). Support-vector networks. Mach Learn 20(3): 273–297

    MATH  Google Scholar 

  3. Davidson I, Fan W (2003) When efficient model averaging out-performs boosting and bagging. In: Proceedings of the 10th European conference on principles and practice of knowledge discovery in databases. Springer, Berlin, pp. 478–486

  4. EPA (1999) Guideline for developing an ozone forecasting program. EPA-454/R-99-009

  5. Fan W, Davidson I (2006) ReverseTesting: an efficient framework to select amongst classifiers under sample selection bias. In: Proceedings of the 12th ACM SIGKDD international conference on knowledge discovery and data mining. Philadelphia

  6. Fan W, Wang H, Yu P, Ma S (2003) Is random model better? On its accuracy and efficiency. In: Proceedings of the 3rd IEEE international conference on data mining

  7. Ferri C, Flach P and Hernndez J (2003). Decision trees for ranking: effect of new smoothing methods, new splitting criteria and simple pruning methods. Technical report, UPV(DSIC 2003)

    Google Scholar 

  8. Forswall CD, Higgins KE (2006) Clean air act implementation in Houston: an historical perspective, 1970–2005 Technical report, Rice University, Environmental and Energy Systems Institute, Shell Center for Sustainability

  9. Ghiaus C (2005). Linear fuzzy-discriminant analysis applied to forecast ozone concentration classes in sea-breeze regime. Atmos Environ 39(26): 4691–4702

    Article  Google Scholar 

  10. Janssen N, Sanderson E (2004) Air-pollution exposure assessment. http://airnet.iras.uu.nl

  11. Kim Y and Kim J (2006). Convex Hull ensemble machine for regression and classification. Knowledge Info Sys 6(6): 645–663

    Article  Google Scholar 

  12. Lambeth B (2006) Ozone maximum model forecast version. In: Proceedings of the national air quality conference, San Antonio

  13. Ling CX, Yan J (2003) Decision tree with better ranking. In: The Proceedings of the 20th international conference on machine learning

  14. Mamitsuka H (2006). Query-learning-based iterative feature-subset selection for learning from high-dimensional data sets. Knowl Info Syst 9(1): 91–108

    Article  Google Scholar 

  15. McMillan N, Bortnicka S, Irwinb M and Berlinerc LM (2005). A hierarchical Bayesian model to estimate and forecast ozone through space and time. Atmos Environ 39(8): 1373–1382

    Article  Google Scholar 

  16. Mintz R, Young B and Svrcek W (2005). Fuzzy logic modeling of surface ozone concentrations. Comput Chem Eng 29(10): 2049–2059

    Article  Google Scholar 

  17. Mitchell T (1997) Machine learning. McGraw Hill

  18. NCDC (2000) http://www.ncdc.noaa.gov/oa/ncdc.html

  19. Ortega S, Soler MR, Beneito J, Pino D (2004) Evaluating of two ozone air quality modeling systems. Atmos Chem Phys Discussi 4:1855–1885, European Geosciences Union

    Google Scholar 

  20. Provost F and Domingos P (2003). Tree induction for probability-based rankings. Mach Learn 52(3): 199–215

    Article  MATH  Google Scholar 

  21. Schlink U, Dorlingb S, Pelikanc E, Nunnarid G, Cawleye G, Junninenf H, Greigg A, Foxallb R, Ebenc K, Chattertonb T, Vondracekc J, Richtera M, Dostalc M, Bertuccod L, Kolehmainenf M and Doyleb M (2003). A rigorous inter-comparison of ground-level ozone predictions. Atmos Environ 37: 3237–3253

    Article  Google Scholar 

  22. Wu X, Yu P, Piatetsky-Shapiro G, Cercone N, Lin TY, Kotagiri R and Wah BW (2003). Data mining: how research meets practical development?. Knowl Info Syst 5(2): 248–261

    Article  Google Scholar 

  23. Zadrozny B (2004) Learning and evaluating classifiers under sample selection bias. In: Proceedings of the 21st international conference on machine learning. Morgan Kaufmann, Sanfransisco

  24. Zhang K, Xu Z, Peng J, Buckles B (2005) Learning through changes: an empirical study of dynamic behaviors of probability estimation trees. In: Proceedings of the 5th IEEE international conference on data mining

Download references

Author information

Authors and Affiliations

  1. Department of Computer Science, Xavier University, New Orleans, LA, USA

    Kun Zhang

  2. IBM T.J.Watson Research, Hawthorne, NY, USA

    Wei Fan

Authors
  1. Kun Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Wei Fan
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Wei Fan.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Zhang, K., Fan, W. Forecasting skewed biased stochastic ozone days: analyses, solutions and beyond. Knowl Inf Syst 14, 299–326 (2008). https://doi.org/10.1007/s10115-007-0095-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10115-007-0095-1

Keywords

Search

Navigation