Time Series
Recently Published Documents


(FIVE YEARS 36780)



Jialing Xu ◽  
Jingxing He ◽  
Jinqiang Gu ◽  
Huayang Wu ◽  
Lei Wang ◽  

Considering the problems of the model collapse and the low forecast precision in predicting the financial time series of the generative adversarial networks (GAN), we apply the WGAN-GP model to solve the gradient collapse. Extreme gradient boosting (XGBoost) is used for feature extraction to improve prediction accuracy. Alibaba stock is taken as the research object, using XGBoost to optimize its characteristic factors, and training the optimized characteristic variables with WGAN-GP. We compare the prediction results of WGAN-GP model and classical time series prediction models, long short term memory (LSTM) and gate recurrent unit (GRU). In the experimental stage, root mean square error (RMSE) is chosen as the evaluation index. The results of different models show that the RMSE of WGAN-GP model is the smallest, which are 61.94% and 47.42%, lower than that of LSTM model and GRU model respectively. At the same time, the stock price data of Google and Amazon confirm the stability of WGAN-GP model. WGAN-GP model can obtain higher prediction accuracy than the classical time series prediction model.

Jidong Yuan ◽  
Mohan Shi ◽  
Zhihai Wang ◽  
Haiyang Liu ◽  
Jinyang Li

2022 ◽  
Matt N Williams ◽  
Stephen Robert Hill

This letter responds to another letter in the journal (Thorney et al., 2021). Thornley et al. reported a time series analysis of "parasuicide" cases in New Zealand children and concluded that this time series demonstrates a "clear detrimental effect of COVID-19 lockdown policies on child mental health”. In this letter, we point out four reasons why this conclusion is not justified by the data presented by Thornley et al. [Note: Our letter did not have a formal abstract; this summary was added for the preprint.]

2022 ◽  
Vol 12 (1) ◽  
Yukinari Seshimo ◽  
Shoichi Yoshioka

AbstractLong-term slow slip events (L-SSEs) have repeatedly occurred beneath the Bungo Channel in southwestern Japan with durations of several months to a couple of years, with a recurrence interval of approximately 6 years. We estimated the spatiotemporal slip distributions of the 2018–2019 Bungo Channel L-SSE by inverting processed GNSS time series data. This event was divided into two subevents, with the first on the southwest side of the Bungo Channel from 2018.3 to 2018.7 and the second beneath the Bungo Channel from 2018.8 to 2019.4. Tectonic tremors became active on the downdip side of the L-SSE occurrence region when large slow slips took place beneath the Bungo Channel. Compared with the previous Bungo Channel L-SSEs, this spatiotemporal slip pattern and amount were similar to those of the 2002–2004 L-SSE. However, the slip expanded in the northeast and southwest directions in the latter half of the second subevent. The maximum amount of slip, the maximum slip velocity, the total released seismic moment, and the moment magnitude of the 2018–2019 L-SSE were estimated to be 28 cm, 54 cm/year, $$4.4 \times 10^{19}$$ 4.4 × 10 19 Nm, and 7.0, respectively, all of which were the largest among the 1996–1998, 2002–2004, 2009–2011, and 2018–2019 L-SSEs.

Pratik Saha ◽  
Pritthijit Nath ◽  
Asif Iqbal Middya ◽  
Sarbani Roy

2022 ◽  
Vol 29 (1) ◽  
pp. 1-15
Justin Schulte ◽  
Frederick Policelli ◽  
Benjamin Zaitchik

Abstract. Many geophysical time series possess nonlinear characteristics that reflect the underlying physics of the phenomena the time series describe. The nonlinear character of times series can change with time, so it is important to quantify time series nonlinearity without assuming stationarity. A common way of quantifying the time evolution of time series nonlinearity is to compute sliding skewness time series, but it is shown here that such an approach can be misleading when time series contain periodicities. To remedy this deficiency of skewness, a new waveform skewness index is proposed for quantifying local nonlinearities embedded in time series. A waveform skewness spectrum is proposed for determining the frequency components that are contributing to time series waveform skewness. The new methods are applied to the El Niño–Southern Oscillation (ENSO) and the Indian monsoon to test a recently proposed hypothesis that states that changes in the ENSO–Indian monsoon relationship are related to ENSO nonlinearity. We show that the ENSO–Indian rainfall relationship weakens during time periods of high ENSO waveform skewness. The results from two different analyses suggest that the breakdown of the ENSO–Indian monsoon relationship during time periods of high ENSO waveform skewness is related to the more frequent occurrence of strong central Pacific El Niño events, supporting arguments that changes in the ENSO–Indian rainfall relationship are not solely related to noise.

2022 ◽  
Vol 22 (1) ◽  
Michelle T. Pedersen ◽  
Thea O. Andersen ◽  
Amy Clotworthy ◽  
Andreas K. Jensen ◽  
Katrine Strandberg-Larsen ◽  

Abstract Background The COVID-19 pandemic and its associated national lockdowns have been linked to deteriorations in mental health worldwide. A number of studies analysed changes in mental health indicators during the pandemic; however, these studies generally had a small number of timepoints, and focused on the initial months of the pandemic. Furthermore, most studies followed-up the same individuals, resulting in significant loss to follow-up and biased estimates of mental health and its change. Here we report on time trends in key mental health indicators amongst Danish adults over the course of the pandemic (March 2020 - July 2021) focusing on subgroups defined by gender, age, and self-reported previously diagnosed chronic and/or mental illness. Methods We used time-series data collected by Epinion (N=8,261) with 43 timepoints between 20 March 2020 and 22 July 2021. Using a repeated cross-sectional study design, independent sets of individuals were asked to respond to the Copenhagen Corona-Related Mental Health questionnaire at each timepoint, and data was weighted to population proportions. The six mental health indicators examined were loneliness, anxiety, social isolation, quality of life, COVID-19-related worries, and the mental health scale. Gender, age, and the presence of previously diagnosed mental and/or chronic illness were used to stratify the population into subgroups for comparisons. Results Poorer mental health were observed during the strictest phases of the lockdowns, whereas better outcomes occurred during reopening phases. Women, young individuals (<34 yrs), and those with a mental- and/or chronic illness demonstrated poorer mean time-series than others. Those with a pre-existing mental illness further had a less reactive mental health time-series. The greatest differences between women/men and younger/older age groups were observed during the second lockdown. Conclusions People with mental illness have reported disadvantageous but stable levels of mental health indicators during the pandemic thus far, and they seem to be less affected by the factors that result in fluctuating time-series in other subgroups.

2022 ◽  
Vol 5 (1) ◽  
Shun Otsubo ◽  
Sreekanth K. Manikandan ◽  
Takahiro Sagawa ◽  
Supriya Krishnamurthy

AbstractThe rate of entropy production provides a useful quantitative measure of a non-equilibrium system and estimating it directly from time-series data from experiments is highly desirable. Several approaches have been considered for stationary dynamics, some of which are based on a variational characterization of the entropy production rate. However, the issue of obtaining it in the case of non-stationary dynamics remains largely unexplored. Here, we solve this open problem by demonstrating that the variational approaches can be generalized to give the exact value of the entropy production rate even for non-stationary dynamics. On the basis of this result, we develop an efficient algorithm that estimates the entropy production rate continuously in time by using machine learning techniques and validate our numerical estimates using analytically tractable Langevin models in experimentally relevant parameter regimes. Our method only requires time-series data for the system of interest without any prior knowledge of the system’s parameters.

Sign in / Sign up

Export Citation Format

Share Document