scholarly journals Hazard function theory for nonstationary natural hazards

2015 ◽  
Vol 3 (11) ◽  
pp. 6883-6915
Author(s):  
L. K. Read ◽  
R. M. Vogel
Keyword(s):  
Time Series ◽  
Natural Hazards ◽  
Hazard Function ◽  
Failure Time ◽  
Function Analysis ◽  
Exceedance Probability ◽  
Critical Event ◽  
Average Return ◽  
Event Series ◽  
Gp Model

Abstract. Impact from natural hazards is a shared global problem that causes tremendous loss of life and property, economic cost, and damage to the environment. Increasingly, many natural processes show evidence of nonstationary behavior including wind speeds, landslides, wildfires, precipitation, streamflow, sea levels, and earthquakes. Traditional probabilistic analysis of natural hazards based on peaks over threshold (POT) generally assumes stationarity in the magnitudes and arrivals of events, i.e. that the probability of exceedance of some critical event is constant through time. Given increasing evidence of trends in natural hazards, new methods are needed to characterize their probabilistic behavior. The well-developed field of hazard function analysis (HFA) is ideally suited to this problem because its primary goal is to describe changes in the exceedance probability of an event over time. HFA is widely used in medicine, manufacturing, actuarial statistics, reliability engineering, economics, and elsewhere. HFA provides a rich theory to relate the natural hazard event series (X) with its failure time series (T), enabling computation of corresponding average return periods, risk and reliabilities associated with nonstationary event series. This work investigates the suitability of HFA to characterize nonstationary natural hazards whose POT magnitudes are assumed to follow the widely applied Generalized Pareto (GP) model. We derive the hazard function for this case and demonstrate how metrics such as reliability and average return period are impacted by nonstationarity and discuss the implications for planning and design. Our theoretical analysis linking hazard event series X, with corresponding failure time series T, should have application to a wide class of natural hazards with rich opportunities for future extensions.

scholarly journals Hazard function theory for nonstationary natural hazards

2016 ◽  
Vol 16 (4) ◽  
pp. 915-925 ◽  
Author(s):  
Laura K. Read ◽  
Richard M. Vogel
Keyword(s):  
Time Series ◽  
Natural Hazards ◽  
Hazard Function ◽  
Failure Time ◽  
Function Analysis ◽  
Economic Cost ◽  
Exceedance Probability ◽  
Critical Event ◽  
Average Return ◽  
Event Series

Abstract. Impact from natural hazards is a shared global problem that causes tremendous loss of life and property, economic cost, and damage to the environment. Increasingly, many natural processes show evidence of nonstationary behavior including wind speeds, landslides, wildfires, precipitation, streamflow, sea levels, and earthquakes. Traditional probabilistic analysis of natural hazards based on peaks over threshold (POT) generally assumes stationarity in the magnitudes and arrivals of events, i.e., that the probability of exceedance of some critical event is constant through time. Given increasing evidence of trends in natural hazards, new methods are needed to characterize their probabilistic behavior. The well-developed field of hazard function analysis (HFA) is ideally suited to this problem because its primary goal is to describe changes in the exceedance probability of an event over time. HFA is widely used in medicine, manufacturing, actuarial statistics, reliability engineering, economics, and elsewhere. HFA provides a rich theory to relate the natural hazard event series (X) with its failure time series (T), enabling computation of corresponding average return periods, risk, and reliabilities associated with nonstationary event series. This work investigates the suitability of HFA to characterize nonstationary natural hazards whose POT magnitudes are assumed to follow the widely applied generalized Pareto model. We derive the hazard function for this case and demonstrate how metrics such as reliability and average return period are impacted by nonstationarity and discuss the implications for planning and design. Our theoretical analysis linking hazard random variable X with corresponding failure time series T should have application to a wide class of natural hazards with opportunities for future extensions.

Failure time prognosis in manufacturing process using multi-dislocated time series convolutional neural network

2020 ◽  
pp. 095440892097414
Author(s):  
Bohao Li ◽  
Liping Zhao ◽  
Yiyong Yao
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Convolutional Neural Network ◽  
Manufacturing Process ◽  
Failure Time ◽  
Great Difficulty ◽  
Milling Process ◽  
Machining Process ◽  
Massive Data ◽  
The Difference

Failure time prognosis in manufacturing process plays a crucial role in guaranteeing manufacturing safety and reducing maintenance loss. However, most current prognosis methods face great difficulty when handling massive data collected from manufacturing process. Convolutional neural network (CNN) provides an effective way to extract features with massive data. Due to the difference between images and multisensory signals, CNN is not suitable for machining process. Inspired by the idea of CNN, a novel prognosis framework is proposed based on the characteristics of multisensory signals, which is called multi-dislocated time series convolutional neural network (MDTSCNN). The proposed MDTSCNN is composed of multi-dislocate layer, convolutional layer, pooling layer and fully connected layer. By adding a multi-dislocate layer, this model can learn the relationship between different signals and different intervals in periodic multisensory signals. The effectiveness of proposed method is validated by a milling process. Compared to other prognosis method, the proposed MDTSCNN shows enhanced performances in prediction accuracy.

scholarly journals Dissecting Tether’s Nonlinear Dynamics during Covid-19

2020 ◽  
Vol 6 (4) ◽  
pp. 161
Author(s):  
Moinak Maiti ◽  
Zoran Grubisic ◽  
Darko B. Vukovic
Keyword(s):  
Time Series ◽  
Regime Switching ◽  
Return Time ◽  
The Other ◽  
Smooth Transition ◽  
Daily Average ◽  
Star Models ◽  
Average Return ◽  
Threshold Autoregression ◽  
Highly Nonlinear

The present study is on the five cryptocurrency daily mean return time series linearity dynamics during the Covid-19 period. These cryptocurrencies were chosen based on their influence on the market, primarily driven by its market capitalisation. Tether is included as the most important stable coin on the market, nominally pegged to the U.S. dollar (USD). The reason to investigate it is that there are some inconsistencies in its behaviour as opposed to the other four cryptocurrencies. This study found that the behaviour of Tether cryptocurrency daily average return time series pattern is highly nonlinear and chaotic in nature, whereas the other four cryptocurrencies (namely Bitcoin, Ethereum, XRP and Bitcoin Cash) daily average return time series were found to be linear in nature. To further study Tether’s nonlinear time series rich dynamics, this study deployed one category of the regime switching models popularly known as the threshold regressions. The study estimates fairly suggest that both the threshold autoregression (TAR) and smooth transition autoregressive (STAR) models with lag 1 are adequate to capture the rich nonlinear and chaotic dynamics of Tether’s daily average return time series.

Using AutoRegressive Integrated Moving Average and Gaussian Processes with LSTM neural networks to predict discrete geomagnetic signals

2020 ◽  
Author(s):  
Laurentiu Asimopolos ◽  
Alexandru Stanciu ◽  
Natalia-Silvia Asimopolos ◽  
Bogdan Balea ◽  
Andreea Dinu ◽  
...  
Keyword(s):  
Time Series ◽  
Deep Learning ◽  
Moving Average ◽  
Time Series Forecasting ◽  
Series Data ◽  
Autoregressive Integrated Moving Average ◽  
Frequency Information ◽  
Probabilistic Forecasts ◽  
Geomagnetic Data ◽  
Gp Model

<p>In this paper, we present the results obtained for the geomagnetic data acquired at the Surlari Observatory, located about 30 Km North of Bucharest - Romania. The observatory database contains records from the last seven solar cycles, with different sampling rates.</p><p>We used AR, MA, ARMA and ARIMA (AutoRegressive Integrated Moving Average) type models for time series forecasting and phenomenological extrapolation. ARIMA model is a generalization of an autoregressive moving average (ARMA) model, fitted to time series data to predict future points in the series</p><p>We made spectral analysis using Fourier Transform, that gives us a relevant picture of the frequency spectrum of the signal component, but without locating it in time, while the wavelet analysis provides us with information regarding the time of occurrence of these frequencies. </p><p>Wavelet allows local analysis of magnetic field components through variable frequency windows. Windows with longer time intervals allow us to extract low-frequency information, medium-sized intervals of different sizes lead to medium-frequency information extraction, and very narrow windows highlight the high-frequencies or details of the analysed signals.</p><p>We extend the study of geomagnetic data analysis and predictive modelling by implementing a Long Short-Term Memory (LSTM) recurrent neural network that is capable of modelling long-term dependencies and is suitable for time series forecasting. This method includes a Gaussian process (GP) model in order to obtain probabilistic forecasts based on the LSTM outputs. </p><p>The evaluation of the proposed hybrid model is conducted using the Receiver Operating Characteristic (ROC) Curve that provides a probabilistic forecast of geomagnetic storm events. </p><p>In addition, reliability diagrams are provided in order to support the analysis of the probabilistic forecasting models.</p><p>The implementation of the solution for predicting certain geomagnetic parameters is implemented in the MATLAB language, using the Toolbox Deep Learning Toolbox, which provides a framework for the design and implementation of deep learning models.</p><p>Also, in addition to using the MATLAB environment, the solution can be accessed, modified, or improved in the Jupyter Notebook computing environment.</p>

Sign in / Sign up

Export Citation Format

Share Document