scholarly journals A Study in Weather Model Output Postprocessing: Using the Boosting Method for Thunderstorm Detection

2009 ◽  
Vol 24 (1) ◽  
pp. 211-222 ◽  
Author(s):  
Donat Perler ◽  
Oliver Marchand
Keyword(s):  
Substantial Improvement ◽  
Probability Of Detection ◽  
Model Output ◽  
Adaptive Boosting ◽  
Linear Discriminant ◽  
Output Surface ◽  
Running System ◽  
Weather Model ◽  
Boosting Algorithm ◽  
Boosting Method

Abstract In this work, a new approach to weather model output postprocessing is presented. The adaptive boosting algorithm is used to train a set of simple base classifiers with historical data from weather model output, surface synoptic observation (SYNOP) messages, and lightning data. The resulting overall method then can be used to classify weather model output to identify potential thunderstorms. The method generates a certainty measure between −1 and 1, describing how likely a thunderstorm is to occur. Using a threshold, the measure can be converted to a binary decision. When compared to a linear discriminant and a method currently employed in an expert system from the German Weather Service, boosting achieves the best validation scores. A substantial improvement of the probability of detection of up to 72% and a decrease of the false alarm rate down to 34% can be achieved for the identification of thunderstorms in model analysis. Independent of the verification results, the method has several useful properties: good cross-validation results, short learning time (≤10 min sequential run time for the experiments on a standard PC), comprehensible inner values of the underlying statistical analysis, and the simplicity of adding predictors to a running system. This paper concludes with a set of possible other applications and extensions to the presented example of thunderstorm detection.

scholarly journals Noisy Reverberation Suppression Using AdaBoost Based EMD in Underwater Scenario

2014 ◽  
Vol 2014 ◽  
pp. 1-7 ◽  
Author(s):  
Kusma Kumari Cheepurupalli ◽  
Raja Rajeswari Konduri
Keyword(s):  
Signal Processing ◽  
Acoustic Signal ◽  
Probability Of Detection ◽  
Chirp Signal ◽  
Processing Scheme ◽  
Adaptive Boosting ◽  
Mode Decomposition ◽  
Crucial Problem ◽  
Filtering Technique ◽  
Simulation Results

Reverberation suppression is a crucial problem in sonar communications. If the acoustic signal is radiated in the water as medium then the degradation is caused due to the reflection coming from surface, bottom, and volume of water. This paper presents a novel signal processing scheme that offers an improved solution in reducing the effect of interference caused due to reverberation. It is based on the combination of empirical mode decomposition (EMD) and adaptive boosting (AdaBoost) techniques. AdaBoost based EMD filtering technique is used for reverberation corrupted chirp signal to decrease the noisy components present in the received signal. An improvement in the probability of detection is achieved using the proposed algorithm. The simulation results are obtained for various reverberation times at various SNR levels.

Domain adaptive boosting method and its applications

2015 ◽  
Vol 24 (2) ◽  
pp. 023038 ◽  
Author(s):  
Jie Geng ◽  
Zhenjiang Miao
Keyword(s):  
Adaptive Boosting ◽  
Boosting Method

scholarly journals Identification of Southern Ocean acoustic targets using aggregation backscatter and shape characteristics

2003 ◽  
Vol 60 (3) ◽  
pp. 641-649 ◽  
Author(s):  
Rachel S Woodd-Walker ◽  
Jonathan L Watkins ◽  
Andrew S Brierley
Keyword(s):  
Antarctic Krill ◽  
Target Identification ◽  
Function Analysis ◽  
Substantial Improvement ◽  
Acoustic Energy ◽  
Biomass Estimation ◽  
Accurate Identification ◽  
Linear Discriminant ◽  
Energy Parameters ◽  
Target Fishing

Abstract Acoustic surveys for biomass estimation require accurate identification of echoes from the target species. In one objective technique for identifying Antarctic krill, the difference between mean volume-backscattering strength at two frequencies is used, but can misclassify small krill and other plankton. Here, we investigate ways to improve target identification by including characteristics of backscattering energy and morphology of aggregations. To do this, multi-frequency acoustic data were collected concurrently with target fishing of Antarctic krill and other euphausiid and salp aggregations. Parameter sets for these known aggregations were collated and used to develop empirical classifications. Both linear discriminant-function analysis (DFA) and the artificial neural network technique were employed. In both cases, acoustic-backscattering energy parameters were most important for discriminating between Antarctic krill and other zooplankton. However, swarm morphology and other parameters improved the discrimination, particularly between krill and salps. Our study suggests that for krill-biomass estimates, a simple DFA based on acoustic-energy parameters is a substantial improvement over current dB-difference acoustic methods; but studies requiring the discrimination of zooplankton other than krill must still be supported by target fishing.

scholarly journals Classification of Disaster Risks in the Philippines using Adaptive Boosting Algorithm with Decision Trees and Support Vector Machine as Based Estimators

2021 ◽  
Vol 4 (1) ◽  
pp. 7-18
Author(s):  
Donata D Acula
Keyword(s):  
Support Vector Machine ◽  
Decision Trees ◽  
Ensemble Methods ◽  
The Philippines ◽  
Support Vector ◽  
Adaptive Boosting ◽  
Risk Levels ◽  
Disaster Data ◽  
Boosting Algorithm ◽  
Intelligent Approach

This paper employed the intelligent approach based on machine learning categorized as base and ensemble methods in classifying the disaster risk in the Philippines. It focused on the Decision Trees, Support Vector Machine, Adaptive Boosting Algorithm with Decision Trees, and Support Vector Machine as base estimators. The research used the Exponential Regression for missing value imputation and converted the number of casualties, damaged houses, and properties into five (5) risk levels using Quantile Method. The 10-fold cross-validation was used to validate the proposed algorithms. The experiment shows that Decision Trees and Adaptive Decision Trees are the most suitable models for the disaster data with the score of more than 90%, more than 75%, more than  75%  in all the classification metrics (accuracy, precision, recall f1-score) when applied to classification risk levels of casualties, damaged houses and damaged properties respectively.

Sign in / Sign up

Export Citation Format

Share Document