A Neural Network for Detecting and Diagnosing Tornadic Circulations

Slides:



Advertisements
Similar presentations
The use of Doppler radar products for nowcasting tornadoes DWD workshop on tornado forecasting, February 24th 2005, Langen 1. introduction 2. products.
Advertisements

Brief introduction on Logistic Regression
Radar Climatology of Tornadoes in High Shear, Low CAPE Environments in the Mid-Atlantic and Southeast Jason Davis Matthew Parker North Carolina State University.
Quantification of Spatially Distributed Errors of Precipitation Rates and Types from the TRMM Precipitation Radar 2A25 (the latest successive V6 and V7)
CSCI 347 / CS 4206: Data Mining Module 07: Implementations Topic 03: Linear Models.
Supervised Learning Recap
Combining Inductive and Analytical Learning Ch 12. in Machine Learning Tom M. Mitchell 고려대학교 자연어처리 연구실 한 경 수
Ounding nalog etrieval ystem Ryan Jewell Storm Prediction Center Norman, OK SARS Sounding Analog Retrieval System.
Sparse vs. Ensemble Approaches to Supervised Learning
1 Learning to Detect Objects in Images via a Sparse, Part-Based Representation S. Agarwal, A. Awan and D. Roth IEEE Transactions on Pattern Analysis and.
Neural Networks Marco Loog.
Neural Networks. R & G Chapter Feed-Forward Neural Networks otherwise known as The Multi-layer Perceptron or The Back-Propagation Neural Network.
Tornado Detection Algorithm (TDA) By: Jeffrey Curtis and Jessica McLaughlin.
Sparse vs. Ensemble Approaches to Supervised Learning
Data mining and statistical learning - lecture 11 Neural networks - a model class providing a joint framework for prediction and classification  Relationship.
CHAPTER 12 ADVANCED INTELLIGENT SYSTEMS © 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang.
Machine Learning1 Machine Learning: Summary Greg Grudic CSCI-4830.
Artificial Neural Networks (ANN). Output Y is 1 if at least two of the three inputs are equal to 1.
Chapter 9 Neural Network.
Chapter 11 – Neural Networks COMP 540 4/17/2007 Derek Singer.
Outline What Neural Networks are and why they are desirable Historical background Applications Strengths neural networks and advantages Status N.N and.
Ensemble Methods: Bagging and Boosting
NSF Medium ITR Real-Time Mining of Integrated Weather Information Setup meeting (Aug. 30, 2002)
GENDER AND AGE RECOGNITION FOR VIDEO ANALYTICS SOLUTION PRESENTED BY: SUBHASH REDDY JOLAPURAM.
Neural Networks Presented by M. Abbasi Course lecturer: Dr.Tohidkhah.
1 Validation for CRR (PGE05) NWC SAF PAR Workshop October 2005 Madrid, Spain A. Rodríguez.
Discriminative Training and Machine Learning Approaches Machine Learning Lab, Dept. of CSIE, NCKU Chih-Pin Liao.
Diagnostic verification and extremes: 1 st Breakout Discussed the need for toolkit to build beyond current capabilities (e.g., NCEP) Identified (and began.
NEXRAD Data Quality 25 August 2000 Briefing Boulder, CO Cathy Kessinger Scott Ellis Joe VanAndel Don Ferraro Jeff Keeler.
Eco 6380 Predictive Analytics For Economists Spring 2016 Professor Tom Fomby Department of Economics SMU.
Neural networks (2) Reminder Avoiding overfitting Deep neural network Brief summary of supervised learning methods.
1 Application of MET for the Verification of the NWP Cloud and Precipitation Products using A-Train Satellite Observations Paul A. Kucera, Courtney Weeks,
Data Mining: Concepts and Techniques1 Prediction Prediction vs. classification Classification predicts categorical class label Prediction predicts continuous-valued.
Machine Learning Usman Roshan Dept. of Computer Science NJIT.
CMPS 142/242 Review Section Fall 2011 Adapted from Lecture Slides.
Travis Smith U. Of Oklahoma & National Severe Storms Laboratory Severe Convection and Climate Workshop 14 Mar 2013 The Multi-Year Reanalysis of Remotely.
Combining Models Foundations of Algorithms and Machine Learning (CS60020), IIT KGP, 2017: Indrajit Bhattacharya.
Big data classification using neural network
High Resolution Weather Radar Through Pulse Compression
Chapter 7. Classification and Prediction
Deep Feedforward Networks
Summary of “Efficient Deep Learning for Stereo Matching”
Paper Review Jennie Bukowski ATS APR-2017
Boosting and Additive Trees
GOES-R Risk Reduction Research on Satellite-Derived Overshooting Tops
A Real-Time Automated Method to Determine Forecast Confidence Associated with Tornado Warnings Using Spring 2008 NWS Tornado Warnings John Cintineo Cornell.
Support Vector Machines (SVM)
Data Mining Lecture 11.
Baselining PMU Data to Find Patterns and Anomalies
Nic Wilson’s M.S.P.M. Research
Tomasz Maszczyk and Włodzisław Duch Department of Informatics,
ECE 471/571 - Lecture 17 Back Propagation.
Automated Extraction of Storm Characteristics
Tornado Warning Verification and its Relationship to Storm Type
Chapter 12 Advanced Intelligent Systems
A Real-Time Learning Technique to Predict Cloud-To-Ground Lightning
Camera Calibration Using Neural Network for Image-Based Soil Deformation Measurement Systems Zhao, Honghua Ge, Louis Civil, Architectural, and Environmental.
Artificial Neural Networks
Cluster Validity For supervised classification we have a variety of measures to evaluate how good our model is Accuracy, precision, recall For cluster.
Aiding Severe Weather Forecasting
Ensemble learning.
Product moment correlation
Model generalization Brief summary of methods
Neural networks (1) Traditional multi-layer perceptrons
A New Approach to Tornado Warning Guidance Algorithms
Identifying Severe Weather Radar Characteristics
David Kauchak CS158 – Spring 2019
A Data Partitioning Scheme for Spatial Regression
Advisor: Dr.vahidipour Zahra salimian Shaghayegh jalali Dec 2017
Presentation transcript:

A Neural Network for Detecting and Diagnosing Tornadic Circulations V Lakshmanan, Gregory Stumpf, Arthur Witt University of Oklahoma, National Severe Storms Laboratory, Meteorological Development Laboratory 2/17/2019 lakshman@ou.edu

Motivation MDA and NSE developed at NSSL MDA identifies storm-scale circulations Which may be precursors to tornadoes Marzban (1997) developed a NN based on MDA parameters to classify tornadoes Using 43 cases Found incorporation of NSE promising Radar Operations Center wanted us to examine using a MDA+NSE NN operationally. Extended Marzban’s work to 83 cases With a few modifications 2/17/2019 lakshman@ou.edu

MDA and NSE Mesocyclone Detection Algorithm (MDA) designed to detect a wide variety of circulations of varying size and strength by analyzing the radial velocity data from a Doppler weather radar 23 attributes for each circulation Near Storm Environment (NSE) Uses analysis grids from the RUC model to derive 245 different attributes. Full list of attributes used is in the conference pre-prints. 2/17/2019 lakshman@ou.edu

Scalar Measures of performance POD = hit / (hit + miss) FAR = fa / (hit + fa) CSI = hit / (hit + miss + fa) HSS = 2*(null * hit - miss * fa) / {(fa+hit)*(fa+null) + (null + miss)*(miss + hit)} We also report Receiver Operating Characteristic (ROC curves) 2/17/2019 lakshman@ou.edu

Neural Network Fully feedforward resilient backpropagation NN Tanh activation function on hidden nodes Logistic (sigmoid) activiation function on output node Error function: weighted sum of cross-entropy and squared sum of all the weights in the network (weight decay) 2/17/2019 lakshman@ou.edu

Truthing Ground truth based on temporal and spatial promixity Done by hand: every circulation was classified. Look for radar signature 20 minutes before a tornado is on the ground to 5 minutes after. 2/17/2019 lakshman@ou.edu

NN Training Method Extract out truthed MDA detections Normalize the input features Determine apriori probability thresholds 13 attributes known to have univariate tendencies and prune the training set Divide set in the ratio 46:20:34 (train: validate: test) Bootstrap train/validate sets. 2/17/2019 lakshman@ou.edu

NN training method (contd.) Find optimal number of hidden nodes Beyond which validation cross-entropy error increases Choose as warning threshold the threshold at which the output of NN on validation set has maximum HSS. 2/17/2019 lakshman@ou.edu

Our method vs. Marzban and Stumpf Slightly different from Marzban/Stumpf: Error criterion different Weight decay Error minimization method different RProp vs SCG Bootstrapped case-wise instead of pattern-wise Automatic pruning based on apriori prob. 2/17/2019 lakshman@ou.edu

43-case comparison Method POD FAR CSI HSS Marzban 0.36 0.69 0.20 0.29 So, we compared against the same 43-cases (with same independent test cases) Most of the difference due to better generalization case-wise bootstrapping Method POD FAR CSI HSS Marzban 0.36 0.69 0.20 0.29 Us 0.34 0.38 0.28 2/17/2019 lakshman@ou.edu

MDA NN (83 case) 43 case data set used by Marzban were large/tall/strong Rather easy dataset of tornado detection The next 40 cases more atypical Mini-supercells, squall-line tornadoes, tropical events etc. Manually selected independent 27 cases to have similar distribution of strong and weak tornadoes. Remaining 56 cases used to verify network. Then, use all 83 cases to create “operational” network. 2/17/2019 lakshman@ou.edu

83 case MDA NN Method POD FAR CSI HSS Test 27 0.44 0.53 0.29 0.41 The performance of best network on independent test case of 27 compared with results on 43-case. And performance of best network trained using all 83 cases (no independent test case) Method POD FAR CSI HSS Test 27 0.44 0.53 0.29 0.41 43-case 0.34 0.38 0.28 Val. (83) 0.42 0.51 0.40 2/17/2019 lakshman@ou.edu

83 case MDA NN ROC curves for 27-case independent test 2/17/2019 lakshman@ou.edu

MDA + NSE Statistics of the dataset change dramatically when we add NSE parameters as inputs 10x as many inputs, so chances of over-fitting much greater. NSE parameters not tied to individual detections NSE parameters highly correlated in space and time. NSE parameters not resolved to radar resolution (20kmx20km vs. 1kmx1km) NSE parameters available hourly; radar data every 5-6 minutes. 2/17/2019 lakshman@ou.edu

Feature Selection Reduce parameters from 245 to 76 based on meteorological understanding. Remove one attribute of highly correlated pairs (Pearson’s correlation coefficient). Take the top “f” fraction of univariate predictors 2/17/2019 lakshman@ou.edu

Choose most general network Variation of the neural network training and validation errors as the number of input features is increased. Choose the number of features where generalization error is minimum (f=0.3) 2/17/2019 lakshman@ou.edu

MDA+NSE On independent 27-case set. Inputs POD FAR CSI HSS MDA 0.44 0.53 0.29 0.41 MDA+NSE 0.47 0.49 0.32 0.45 2/17/2019 lakshman@ou.edu

MDA+NSE (27-case set) 2/17/2019 lakshman@ou.edu

Generalization Similar HSS scores on training, validation and independent test data sets. In MDA+NSE, we sacrificed higher performance to get better generalization Inputs POD FAR CSI HSS MDA 0.44 0.53 0.29 0.41 MDA+NSE 0.47 0.49 0.32 0.45 2/17/2019 lakshman@ou.edu

Is NSE information helpful? NSE parameters changed the statistics of the data set The MDA+NSE neural network is only marginally better than a MDA NN but: NSE information has the potential to be useful. We used only 4 of the 76 of the 245 features! Inputs POD FAR CSI HSS MDA 0.44 0.53 0.29 0.41 MDA+NSE 0.47 0.49 0.32 0.45 2/17/2019 lakshman@ou.edu

Going further Where can we go further with this approach? Find better ways to reduce the number of features Use time history of detections Generate many more data cases. All of which will yield very little (we believe). 2/17/2019 lakshman@ou.edu

Spatio-temporal Tornado Guidance Formulate the tornado prediction problem differently. Instead of devising a machine intelligence approach to classify detections Spatio-temporal: of estimating the probability of a tornado event at a particular spatial location within a given time window 2/17/2019 lakshman@ou.edu

Spatio-temporal approach Our initial approach: Modify ground truth to create spatial truth field use a least-squares methodology to estimate shear morphological image processing to estimate gradients, fuzzy logic to generate compact measures of tornado possibility a classification neural network to generate the final spatio-temporal probability field. Past and future history, both of observed tornadoes and of the candidate regions, is obtained by tracking clustered radar reflectivity values integrate data from other sensors (e.g: numerical models and lightning). Paper at the IJCNN 2005 2/17/2019 lakshman@ou.edu

Acknowledgements Funding for this research was provided under NOAA-OU Cooperative Agreement NA17RJ1227 and supported by the Radar Operations Center. Caren Marzban and Don Burgess, both of the University of Oklahoma, helped us immensely on the methods and attributes used in this paper 2/17/2019 lakshman@ou.edu