TITLE

Regularized Logistic Models for Probabilistic Forecasting and Diagnostics

AUTHOR(S)
Bröcker, Jochen
PUB. DATE
February 2010
SOURCE
Monthly Weather Review;Feb2010, Vol. 138 Issue 2, p592
SOURCE TYPE
Academic Journal
DOC. TYPE
Article
ABSTRACT
Logistic models are studied as a tool to convert dynamical forecast information (deterministic and ensemble) into probability forecasts. A logistic model is obtained by setting the logarithmic odds ratio equal to a linear combination of the inputs. As with any statistical model, logistic models will suffer from overfitting if the number of inputs is comparable to the number of forecast instances. Computational approaches to avoid overfitting by regularization are discussed, and efficient techniques for model assessment and selection are presented. A logit version of the lasso (originally a linear regression technique), is discussed. In lasso models, less important inputs are identified and the corresponding coefficient is set to zero, providing an efficient and automatic model reduction procedure. For the same reason, lasso models are particularly appealing for diagnostic purposes.
ACCESSION #
48596211

 

Related Articles

  • Conditional Exceedance Probabilities. Mason, Simon J.; Galpin, Jacqueline S.; Goddard, Lisa; Graham, Nicholas E.; Rajartnam, Balakanapathy // Monthly Weather Review;Feb2007, Vol. 135 Issue 2, p363 

    Probabilistic forecasts of variables measured on a categorical or ordinal scale, such as precipitation occurrence or temperatures exceeding a threshold, are typically verified by comparing the relative frequency with which the target event occurs given different levels of forecast confidence....

  • A generalized Dantzig selector with shrinkage tuning. Gareth M. James; Peter Radchenko // Biometrika;Jun2009, Vol. 96 Issue 2, p323 

    The Dantzig selector performs variable selection and model fitting in linear regression. It uses an L1 penalty to shrink the regression coefficients towards zero, in a similar fashion to the lasso. While both the lasso and Dantzig selector potentially do a good job of selecting the correct...

  • Statistical Prediction of Weekly Tropical Cyclone Activity in the Southern Hemisphere. Leroy, Anne; Wheeler, Matthew C. // Monthly Weather Review;Oct2008, Vol. 136 Issue 10, p3637 

    A statistical prediction scheme, employing logistic regression, is developed to predict the probability of tropical cyclone (TC) formation in zones of the Southern Hemisphere during forthcoming weeks. Through physical reasoning, examination of previous research, and some new analysis, five...

  • A Multimodel Analysis, Validation, and Transferability Study of Global Soil Wetness Products. Gao, Xiang; Dirmeyer, Paul A. // Journal of Hydrometeorology;Dec2006, Vol. 7 Issue 6, p1218 

    Multimodel ensemble forecasting has been shown to offer a systematic improvement in the skill of climate prediction with atmosphere and ocean circulation models. However, little such work has been done for the land surface component, an important lower boundary for weather and climate forecast...

  • A Perfect Prognosis Scheme for Forecasting Warm-Season Lightning over Florida. Shafer, Phillip E.; Fuelberg, Henry E. // Monthly Weather Review;Jun2008, Vol. 136 Issue 6, p1817 

    This study develops and evaluates a statistical scheme for forecasting warm-season lightning over Florida. Four warm seasons of analysis data from the Rapid Update Cycle (RUC) and lightning data from the National Lightning Detection Network are used in a perfect prognosis technique to develop a...

  • Development of a perfect prognosis probabilistic model for prediction of lightning over south-east India. Rajeevan, M; Madhulatha, A; Rajasekhar, M; Bhate, Jyoti; Kesarkar, Amit; Rao, B // Journal of Earth System Science;Apr2012, Vol. 121 Issue 2, p355 

    A prediction model based on the perfect prognosis method was developed to predict the probability of lightning and probable time of its occurrence over the south-east Indian region. In the perfect prognosis method, statistical relationships are established using past observed data. For real time...

  • Long-range forecasting of intermittent streamflow. van Ogtrop, F. F.; Vervoort, R. W.; Heller, G. Z.; Stasinopoulos, D. M.; Rigby, R. A. // Hydrology & Earth System Sciences Discussions;2011, Vol. 8 Issue 1, p681 

    Long-range forecasting of intermittent streamflow in semi-arid Australia poses a number of major challenges. One of the challenges relates to modelling zero, skewed, non-stationary, and non-linear data. To address this, a probabilistic statistical model to forecast streamflow 12 months ahead is...

  • Robust methods for the analysis of spatially autocorrelated data. Cerioli, Andrea; Riani, Marco // Statistical Methods & Applications;2002, Vol. 11 Issue 3, p335 

    Proposes a robust technique for the analysis of spatial data through simultaneous autoregressive models. Number of useful diagnostics which are monitored along the search to help identify masked spatial outliers and high leverage sites; Introduction to Forward Search, a developed tool for...

  • Graphics for studying logistic regression models. Scrucca, Luca // Statistical Methods & Applications;2002, Vol. 11 Issue 3, p371 

    Focuses on logistic regression models for binary responses. Different methods for studying the log-density ratio through graphs, which allows the selection which predictors are needed; Discussion on data analysis examples based on real datasets available in literature.

Share

Read the Article

Courtesy of THE LIBRARY OF VIRGINIA

Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics