Search Results (5)

This study proposes a hybrid approach to analyze and forecast non-stationary financial time series by combining statistical models with deep neural networks. A model is introduced that integrates three key components: the Generalized Autoregressive Score (GAS) model, which captures volatility dynamics; an attention mechanism (ATT), which identifies the most relevant features within the sequence; and a Long Short-Term Memory (LSTM) neural network, which receives the outputs of the previous modules to generate price forecasts. This architecture is referred to as GAS-ATT-LSTM. Both unidirectional and bidirectional variants were evaluated using real financial data from the Nasdaq Composite Index, Invesco QQQ Trust, ProShares UltraPro QQQ, Bitcoin, and gold and silver futures. The proposed model’s performance was compared against five benchmark architectures: LSTM Bidirectional, GARCH-LSTM Bidirectional, ATT-LSTM, GAS-LSTM, and GAS-LSTM Bidirectional, under sliding windows of 3, 5, and 7 days. The results show that GAS-ATT-LSTM, particularly in its bidirectional form, consistently outperforms the benchmark models across most assets and forecasting horizons. It stands out for its adaptability to varying volatility levels and temporal structures, achieving significant improvements in both accuracy and stability. These findings confirm the effectiveness of the proposed hybrid model as a robust tool for forecasting complex financial time series. Full article

Timely diagnosis and prognosis based on degradation symptoms are essential steps for condition-based maintenance (CBM) to guarantee industrial safety and productivity. Most industrial machines operate under variable operating conditions. This time-varying operating condition can accelerate the machinery’s degradation process. It may have a massive influence on data and impede the process of diagnosis and prognosis of the machinery. Therefore, in this paper, to address the mentioned problems, we introduced an approach for modelling non-stationary long-term condition monitoring data. This procedure includes separating random and deterministic parts and identifying possible autodependence hidden in the random sequence, as well as potential time-dependent variance. To achieve these objectives, we employ a time-varying coefficient autoregressive (TVC-AR) model within a Bayesian framework. However, due to the limited availability of diverse run-to-failure data sets, we validate the proposed procedure using a simulated degradation model and two widely recognized benchmark data sets (FEMTO and wind turbine drive), which demonstrate the model’s effectiveness in capturing complex non-stationary degradation characteristics. Full article

Underwater signal processing is primarily based on sound waves because of the unique properties of water. However, the slow speed and limited bandwidth of sound introduce numerous challenges, including pronounced time-varying characteristics and significant multipath effects. This paper explores a channel estimation method utilizing superimposed training sequences. Compared with conventional schemes, this method offers higher spectral efficiency and better adaptability to time-varying channels owing to its temporal traversal. To ensure success in this scheme, it is crucial to obtain time-varying channel estimation and data detection at low SNRs given that superimposed training sequences consume power resources. To achieve this goal, we initially employ coarse channel estimation utilizing superimposed training sequences. Subsequently, we employ approximate message passing algorithms based on the estimated channels for data detection, followed by iterative channel estimation and equalization based on estimated symbols. We devise an approximate message passing channel estimation method grounded on a Gaussian mixture model and refine its hyperparameters through the expectation maximization algorithm. Then, we refine the channel information based on time correlation by employing an autoregressive hidden Markov model. Lastly, we perform numerical simulations of communication systems by utilizing a time-varying channel toolbox to simulate time-varying channels, and we validate the feasibility of the proposed communication system using experimental field data. Full article

Currently, most of the existing link parameter prediction schemes assume that the link state remains constant during the measurement period, making it difficult to capture their time-varying characteristics. To solve this problem, this paper proposes a prediction problem for packet loss rate in a non-stationary network environment. The measurement period is divided into several adjacent time windows, and the packet loss rates measured passively in each time window are regarded as non-stationary time sequences for real-time tracking to obtain the changes in link packet loss rate at a small cost. By analyzing time-varying autoregressive (TVAR) sequences, a scheme for estimating the time-varying coefficient was presented. In addition, a prediction scheme for the packet loss rate in a non-stationary network was proposed based on TVAR sequences. Finally, this paper conducts experiments based on a non-stationary network simulation environment established by the improved Gilbert–Elliot model and a small wireless multi-hop network experiment platform built in reality. Simulation and experimental results show that the prediction scheme of the packet loss rate based on the TVAR sequence can accurately predict the packet loss rate. Full article

Beat-by-beat variability (BBV) rhythms are observed in both cardiovascular (CV) and intracranial (IC) compartments, yet interactions between the two are not fully understood. Real-Time Phase-Contrast (RT-PC) MRI sequence was acquired for 30 healthy volunteers at 1st cervical level on a 3T scanner. The arterial (AF), venous (VF), and cerebrospinal fluid (CSF) flow (CSFF) were computed as velocity integrals over the internal carotid artery, internal jugular vein, and CSF. AF, VF, and CSFF signals were segmented in inspiration and expiration beats, to assess the respiration influence. Systolic and diastolic BBV, and heart period series underwent autoregressive power spectral density analysis, to evaluate the low-frequency (LF, Mayer waves) and high frequency (HF, respiratory waves) components. The diastolic VF had the largest BBV. LF power was high in the diastolic AF series, poor in all CSFF series. The pulse wave analyses revealed higher mean amplitude during inspiration. Findings suggests a possible role of LF modulation of IC resistances and propagation of HF waves from VF to AF and CCSF. PC-RT-MRI could provide new insight into the interaction between CV and IC regulation and pave the way for a detailed analysis of the cerebrovascular effects of varied respiration patterns due to exercise and rehabilitation. Full article