PAC learning in non-linear FIR models

Najarian, Kayvan; Dumont, Guy A.; Davies, M. S.; Heckman, Nancy E.

doi:10.1002/1099-1115(200102)15:1<37::aid-acs626>3.0.co;2-7

Cited by 15 publications

(6 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SVM was preferred as the general machine learning framework for classification over structures such as neural networks and radial basis function networks, primarily because of studies that have shown that when limited amount of training data is available, neural networks [22] and radial basis functions [23] may not provide desirable generalization performance and may overfit the data.…”

Section: Methodsmentioning

confidence: 99%

Non-linear dynamical signal characterization for prediction of defibrillation success through machine learning

Shandilya

Ward

Kurz

et al. 2012

BMC Med Inform Decis Mak

View full text Add to dashboard Cite

BackgroundVentricular Fibrillation (VF) is a common presenting dysrhythmia in the setting of cardiac arrest whose main treatment is defibrillation through direct current countershock to achieve return of spontaneous circulation. However, often defibrillation is unsuccessful and may even lead to the transition of VF to more nefarious rhythms such as asystole or pulseless electrical activity. Multiple methods have been proposed for predicting defibrillation success based on examination of the VF waveform. To date, however, no analytical technique has been widely accepted. We developed a unique approach of computational VF waveform analysis, with and without addition of the signal of end-tidal carbon dioxide (PetCO2), using advanced machine learning algorithms. We compare these results with those obtained using the Amplitude Spectral Area (AMSA) technique.MethodsA total of 90 pre-countershock ECG signals were analyzed form an accessible preshosptial cardiac arrest database. A unified predictive model, based on signal processing and machine learning, was developed with time-series and dual-tree complex wavelet transform features. Upon selection of correlated variables, a parametrically optimized support vector machine (SVM) model was trained for predicting outcomes on the test sets. Training and testing was performed with nested 10-fold cross validation and 6–10 features for each test fold.ResultsThe integrative model performs real-time, short-term (7.8 second) analysis of the Electrocardiogram (ECG). For a total of 90 signals, 34 successful and 56 unsuccessful defibrillations were classified with an average Accuracy and Receiver Operator Characteristic (ROC) Area Under the Curve (AUC) of 82.2% and 85%, respectively. Incorporation of the end-tidal carbon dioxide signal boosted Accuracy and ROC AUC to 83.3% and 93.8%, respectively, for a smaller dataset containing 48 signals. VF analysis using AMSA resulted in accuracy and ROC AUC of 64.6% and 60.9%, respectively.ConclusionWe report the development and first-use of a nontraditional non-linear method of analyzing the VF ECG signal, yielding high predictive accuracies of defibrillation success. Furthermore, incorporation of features from the PetCO2 signal noticeably increased model robustness. These predictive capabilities should further improve with the availability of a larger database.

show abstract

Section: Methodsmentioning

confidence: 99%

Non-linear dynamical signal characterization for prediction of defibrillation success through machine learning

Shandilya

Ward

Kurz

et al. 2012

BMC Med Inform Decis Mak

View full text Add to dashboard Cite

show abstract

“…To extend the PAC learning theory to a learning scheme for m-dependent sequences with any distribution, we start from an inequality that bounds the summation of a sequence of m-dependent r.v.s. and was introduced in Najarian et al [7].…”

Section: Extension Of Pac Learning To M-dependent Casesmentioning

confidence: 99%

“…The proof of Theorem III.1 given in Najarian et al [7] is inspired by the proof of the central limit theorem for a sequence of dependent data by Iosifescu and coworkers [9]. Notice that if there exists no integer k such that n ϭ k(m ϩ 1), by defining k as k Ϫ [n/(m ϩ 1)], a similar approach can be followed to extend the result of the theorem to such cases.…”

Section: Extension Of Pac Learning To M-dependent Casesmentioning

confidence: 99%

“…As a result, in an NFIR model, the inputs at times t and t ϩ 1 are correlated and consequently dependent. An extension of the PAC learning theory that considers learning with m-dependent uniformly distributed data and includes FIR modeling tasks is given in Najarian et al [6,7]. However, this paradigm of PAC learning with m-dependent is limited to uniformly distributed data and cannot be applied to the cases where the input data is distributed according to any fixed probability.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

FIR Volterra kernel neural models and PAC learning

Najarian

2002

Complexity

View full text Add to dashboard Cite

The probably approximately correct (PAC) learning theory creates a framework to assess the learning properties of static models for which the data are assumed to be independently and identically distributed (i.i.d.). The present article first extends the idea of PAC learning to cover the learning of modeling tasks with m-dependent sequences of data. The data are assumed to be marginally distributed according to a fixed arbitrary probability.The resulting framework is then applied to evaluate learning of Volterra Kernel FIR models.

show abstract

“…A fixed-distribution extension of the PAC learning theory to included FIR models was given in (K. Najarian, Guy A. Dumont, and Michael S. Davies, 2001). The general results of this extension were summarized in a thorem that is briefly described here (for more details see: (K. Najarian, Guy A. Dumont, and Michael S. Davies, 2001) ):…”

Section: Extension Of Pac Learning To M-dependent Casesmentioning

confidence: 99%

Learning of Fir and Arx Neural Networks With Empirical Risk Minimization Algorithm

Najarian

2002

IFAC Proceedings Volumes

View full text Add to dashboard Cite

PAC learning in non-linear FIR models

Cited by 15 publications

References 13 publications

Non-linear dynamical signal characterization for prediction of defibrillation success through machine learning

Non-linear dynamical signal characterization for prediction of defibrillation success through machine learning

FIR Volterra kernel neural models and PAC learning

Learning of Fir and Arx Neural Networks With Empirical Risk Minimization Algorithm

Contact Info

Product

Resources

About