Decomposition Model Research Articles

Prediction of corn futures prices with decomposition and hybrid deep learning models

Prediction of corn futures prices with decomposition and hybrid deep learning models

Assessing Performance of Estimation Techniques in Time Series Analysis when Trend-cycle Component is Linear

Abstract: Two decomposition techniques are Buys-Ballot and least square techniques are presented in this study. The two important patterns that may be discussed are trend and seasonality and two competing models are additive and multiplicative models. The trend-cycle component is linear. The emphasis is to assess the performance of Buys-Ballot estimates and least square estimates using accuracy measures (Mean Error (ME), Mean Square Error (MSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). Results show that the two estimation techniques are very good in estimating the linear trend parameters and seasonal effects when the model for decomposition is additive. It differs for multiplicative model.

Chronic widespread musculoskeletal pain shares a highly heritable latent pathway with atherosclerosis and arterial stiffness

Abstract Chronic widespread pain (CWP) is prevalent and associated with reduced life expectancy. Cardiovascular disease is one possible mechanism for this. The purpose of this study was to examine the association of CWP with arterial stiffness and carotid plaque measured using ultrasound to determine if shared environmental or genetic factors might account for any observed association. Around 3000 participants from the TwinsUK with CWP information and measures of carotid–femoral pulse wave velocity (cfPWV), carotid intima–media thickness (cIMT), and plaque were considered. The relationship between CWP and cfPWV, cIMT, and plaque was determined. UK Biobank data were used to replicate the association. Cholesky decomposition and multivariate pathway twin models were examined. Using a 2-sample Mendelian randomisation approach, the causal association between CWP and coronary artery disease was assessed. TwinsUK participants demonstrated a significant association between CWP and increased cfPWV consistent with arterial stiffening (OR = 1.35, P-value = 0.012), as well as the presence of carotid plaque (OR = 1.45, P-value = 0.8e-5). The twin modelling showed a common latent component and pathway underlying CWP, cfPWV, and carotid plaque, with genetic factors accounting for 68% and 90% of the latent factor variation, respectively. The 2-sample MR revealed a potential causal association between CWP and coronary artery disease. This study found that those with CWP have increased the risk of arterial stiffness and atherosclerosis and suggests that CWP leads to an increased risk of cardiovascular disease through genetic factors.

Detection of the Presence of Seasonal Effect for Some Selected Trending Curves and Choice of Model in Time Series Data

This study discusses detection of the presence of seasonal effect and choice of model in time series data. This study considers the nature of linear, quadratic and exponential trending curves and model structure. The test is constructed in the row and overall variances of the Buys-Ballot table. Empirical example is used to determine the appropriate model for decomposition of the study data.

BS-CP: Efficient streaming Bayesian tensor decomposition method via assumed density filtering.

Tensor data is common in real-world applications, such as recommendation system and air quality monitoring. But such data is often sparse, noisy, and fast produced. CANDECOMP/PARAFAC (CP) is a popular tensor decomposition model, which is both theoretically advantageous and numerically stable. However, learning the CP model in a Bayesian framework, though promising to handle data sparsity and noise, is computationally challenging, especially with fast produced data streams. The fundamental problem addressed by the paper is mainly tackles the efficient processing of streaming tensor data. In this work, we propose BS-CP, a quick and accurate structure to dynamically update the posterior of latent factors when a new observation tensor is received. We first present the BS-CP1 algorithm, which is an efficient implementation using assumed density filtering (ADF). In addition, we propose BS-CP2 algorithm, using Gauss-Laguerre quadrature method to integrate the noise effect which shows better empirical result. We tested BS-CP1 and BS-CP2 on generic real recommendation system datasets, including Beijing-15k, Beijing-20k, MovieLens-1m and Fit Record. Compared with state-of-the-art methods, BS-CP1 achieve 31.8% and 33.3% RMSE improvement in the last two datasets, with a similar trend observed for BS-CP2. This evidence proves that our algorithm has better results on large datasets and is more suitable for real-world scenarios. Compared with most other comparison methods, our approach has demonstrated an improvement of over 10% and exhibits superior stability.

Kinetic model of thermal decomposition of nitric acid solutions of hydrazine nitrate

The kinetic parameters of thermal decomposition of an aqueous solution containing 0.029 mass fractions of hydrazine nitrate and 0.44 mass fractions of nitric acid have been determined: onset temperature and specific thermal effect of exothermic reactions, activation energy, pre-exponential factor, reaction rate and order. Based on the data obtained, a mathematical model of thermal decomposition of nitric acid solutions of hydrazine nitrate has been proposed. It has been found that thermolysis of the solution has a multi-stage nature and can be described by a system of 5 equations of 1st and 2nd order. A comparison of two approaches to creating a mathematical model has been carried out: based on adiabatic calorimetry and differential scanning calorimetry data.

Spatial and temporal analysis of decomposition models in China

This study presents a comprehensive evaluation of 15 decomposition models (12 empirical and 3 atmospheric transmittance models) for estimating diffuse horizontal irradiance at 17 radiation sites in China, using hourly radiation data from 2011 to 2020. Our results show distinct patterns in model performance across different geographical regions, seasons, and sky conditions. The Liu model demonstrates the best overall performance with an RMSE of 78.20 W/m2, while model accuracy shows significant geographical variation, performing best in South and Southeast China (RMSE<70 W/m2) and worst in the Qinghai-Tibet Plateau and Northwest China (RMSE>90 W/m2). Seasonal analysis reveals better performance in winter than in summer, with RMSE differences approaching 40 W/m2, mainly due to the higher proportion of solar elevation angles exceeding 30° in summer. Under different sky conditions (classified by clearness index: 0–0.35 for overcast, 0.35–0.65 for partly cloudy, 0.65–1 for clear skies), most models follow an RMSE pattern of partly cloudy > clear sky > overcast. However, the Reindl2, Boland, DIRINT, and DIRINDEX models deviate from this trend due to their formula structure and sensitivity to atmospheric parameters. To reduce these regional disparities, we propose a new region-specific model selection strategy: the DIRINDEX model for eastern regions, DIRINT for central areas, and Karatasou for western regions. This combined approach reduces the overall RMSE to 73.17 W/m2. This research deepens our understanding of the application of decomposition models in China's complex geographical and climatic conditions, offering valuable references for solar radiation modeling and renewable energy forecasting in diverse climatic regions.

A flexible class of priors for orthonormal matrices with basis function-specific structure

Statistical modeling of high-dimensional matrix-valued data motivates the use of a low-rank representation that simultaneously summarizes key characteristics of the data and enables dimension reduction. Low-rank representations commonly factor the original data into the product of orthonormal basis functions and weights, where each basis function represents an independent feature of the data. However, the basis functions in these factorizations are typically computed using algorithmic methods that cannot quantify uncertainty or account for basis function correlation structure a priori. While there exist Bayesian methods that allow for a common correlation structure across basis functions, empirical examples motivate the need for basis function-specific dependence structure. We propose a prior distribution for orthonormal matrices that can explicitly model basis function-specific structure. The prior is used within a general probabilistic model for singular value decomposition to conduct posterior inference on the basis functions while accounting for measurement error and fixed effects. We discuss how the prior specification can be used for various scenarios and demonstrate favorable model properties through synthetic data examples. Finally, we apply our method to two-meter air temperature data from the Pacific Northwest, enhancing our understanding of the Earth system’s internal variability.

Data-driven forecasting framework for daily reservoir inflow time series considering the flood peaks based on multi-head attention mechanism

Accurate and reliable daily reservoir inflow forecast plays an essential role in several applications involving the management and planning of water resources, such as hydroelectric generation, flood control, water supply, and basin ecological dispatching. Runoff usually exhibits strong non-linearity, high uncertainty, and spatial and temporal variability. Existing techniques fail to capture complete dynamics change processes effectively. A data-driven forecasting framework for daily reservoir inflow time series considering the flood peaks based on a multi-head attention mechanism was developed, referred to as the GWOCS-VMD-CNN-Transformer (GCVCT). First, the model utilize Grey Wolf Optimizer coupled with Cuckoo Search (GWO-CS) algorithms to optimize parameters in variational mode decomposition model (VMD). This approach helps obtain highly correlated intrinsic mode function (IMF) components, enhancing the frequency resolution of the input dataset. The proposed method overcomes the bottleneck of other available methods by decomposing the time series to capture the main long-term and short-term properties of hydrological processes. Second, the convolution neural network and Transformer (CNN-Transformer) are based on a multi-head attention mechanism as the objective predictive method. Finally, six evaluation indicators verify the performance of the proposed approach. The approach’s reliability was evaluated using the historical daily reservoir inflow data from the Xiluodu (XLD) and Wudongde (WDD) reservoirs in the Jinsha River Basin, China. Several single and hybrid models were developed for comparative analysis. The results indicate that the proposed ensemble approach fits better than other developed model methods. The GCVCT model showed excellent performance in forecasting the inflows of XLD and WDD reservoirs, with NSE values of 0.985 and 0.984, respectively. Furthermore, the GCVCT framework forecast capacity for peak inflow was further verified through discussion and analysis of the 48 peak flows during the validation period, consistently outperforming other models in predicting peak flow for both study reservoirs. This framework provides an effective method for the scientific optimal scheduling of hydropower reservoirs, enabling more sustainable and efficient management practices. It also demonstrates the potential of powerful deep-learning models in intelligent hydrological forecasting.

Carbon emission reduction potential of land use in typical alpine meadow region in China

Alpine meadows are primarily situated in mountainous areas and plateaus, exhibiting a high sensitivity to both climate fluctuations and human activities. However, understanding the factors influencing land use carbon emissions change in these regions remains complex, with uncertain mechanisms. Thus, it is imperative to accurately grasp the factors affecting carbon emissions in alpine meadows. We conducted a study estimating land use carbon emissions in the Upper Yellow River in Gannan from 1990 to 2020, using the Logarithmic Mean Divisia Index decomposition model and the Tapio decoupling model to identify driving factors of land use carbon emissions. Additionally, we analyzed the correlation between these factors and economic growth. The findings showed that the land use efficiency in the Upper Yellow River in Gannan region improved substantially by 436 %. Particularly, construction land emerged as the primary source of carbon emissions. The land use carbon emission increased with the increase of carbon emission intensity of construction land and the expansion of the size of construction land. Moreover, economic development and land use carbon emissions are weak decoupled. These results suggest that economic development is considered the primary driving force behind the increase in land use carbon emissions in the region. To achieve the decoupling of economic development and land use carbon emissions, we propose to improve land use efficiency by adjusting the industrial structure as well as improving land use efficiency.

Identification and apportionment of groundwater pollution sources in the Guanzhong region based on PMF model

Identifying pollution sources and measuring their impact is essential for protecting groundwater. Groundwater pollution sources in the Guanzhong region were identified using the positive matrix decomposition (PMF) model, and their contributions were quantified. The spatial distribution of socio-economic factors (population, GDP, fertilizer use, farmland, and livestock) was analyzed. Redundancy and correlation analyses were used to assess the impact of land use on groundwater quality. The results showed that the primary groundwater physicochemical constituents exceeding the standard limits were, in descending order, NO3 −, Na+, TDS, F−, SO4 2−, and Cl−, and the most polluted areas were found in the eastern and southern parts of the study area. Land use and socio-economic factors can have diverse impacts on groundwater quality. Five main sources responsible for groundwater quality deterioration in the Guanzhong were comprehensively identified, i.e., agricultural nonpoint sources, domestic pollution from urban and rural development, natural release of fluorine from stratigraphic minerals as well as coal mine storage and combustion, industrial discharges, and water-rock reactions. Natural release from minerals and industrial discharges are the major causes of groundwater degradation in the study area, which is followed by water-rock reactions, agriculture, and domestic pollution.

Sleep patterns, global mental status and mortality risk among middle-aged urban adults.

Sleep, cognition, and mortality may be interdependent. We explored paths between sleep, cognition and mortality and potential interactions. The study examines the relationship among sleep, global mental status, and mortality risk using data from 1364 participants from the Healthy Aging in Neighborhood of Diversity across the Life Span (HANDLS) study. We used Cox proportional hazards models and four-way decomposition models to analyze sleep patterns and global mental status. After a median time at risk of 8.2 years, 172 deaths occurred, with rate of 16 per 1000 person-years. A 1-unit increase in the Pittsburgh Sleep Quality Index (PSQI) global score was linked to a 7% increase in mortality risk in the reduced model, but this effect was attenuated in the full model. In both reduced and fully adjusted models, the PSQI global score and sleep quality domains interacted with global mental status, with poor sleep generally associated with mortality risk in the group with better global mental status at first-visit. In four-way decomposition models, total effects (TE) of PSQI scores on mortality risk were positive and statistically significant, while being mostly controlled direct effects. However, among women, the inverse TE of global mental status on mortality risk was partially mediated by PSQI sleep latency and the PSQI global. Poor global mental status is associated with greater mortality risk at better sleep quality levels and vice versa. Further longitudinal studies with multiple sleep and cognitive performance repeats are needed to corroborate these findings.

무인자동 양식어류 급이공급 시스템 개발 및 검증

Unmanned automatic feeding systems are being actively distributed to automate land-based aquaculture farms. Currently, feed supply in aquaculture is primarily based on a fixed period and amount, depending on the predetermined feed quantity set by the operator, which does not consider the dynamic feed supply needs of fish. This can lead to under- or oversupply of feed, affecting the fish growth rate and water quality in aquaculture. To address this, a method is proposed in this paper to identify fish hunger levels using machine vision and predict the dissolved oxygen concentration in real time to prevent mass fish mortality due to water quality deterioration, even with additional feed supply. The method aims to verify its performance. Firstly, based on the observation that aquaculture farm operators estimate fish hunger levels by feeding time, a method is proposed to estimate hunger levels using machine vision and feeding time. Next, a combination of time series data decomposition model and pre-trained deep learning model is used to predict the dissolved oxygen concentration in real time.

CoATR: A Convolutional Autoregressive Tensor-Ring decomposition method for sparse spatio-temporal traffic data

Spatio-temporal traffic data collected by various sensing systems have chronic issues of missing and corruption, thus accurate data imputation and prediction have been extensively researched. Previous most methods can be divided into two categories: model-driven methods with physical explanations and data-driven deep learning methods. These methods all achieve promising results due to their respective advantages. However they have some limitations: model-driven methods are often linear and may not accurately model the spatio-temporal complexity, while deep learning methods often lack the physical reality making them probe to over-fitting. Inspired by both, we propose a new Convolutional-based generalized Autoregressive Tensor-Ring decomposition method (CoATR) for the completion of spatio-temporal data. CoATR not only retains the advantages of the tensor-ring (TR) decomposition model for global modeling of spatio-temporal data but also exploits the ability of deep networks to model nonlinear features. To be specific, we introduce TR decomposition to capture the global low-rankness and employ multilayer convolutional neural networks to model the global complex interactions among the TR factors for exploring the nonlinear features of the spatio-temporal data. Moreover, we design a new autoregressive network to further explore the local temporal variation in the data. Extensive experiments on a variety of common traffic datasets have validated the effectiveness and superiority of the CoATR over classical model-driven methods and other state-of-the-art data-driven deep learning methods.

Decomposition framework for long term load forecasting on temperature insensitive area

Long-term load forecasting (LTLF) with hourly resolution provides more information for power system planning and becomes increasingly important as the growing penetration of renewable energy sources intensifies system uncertainties. In most LTLF literature, temperature is used as the main driving factor. However, the system load in an area with a high proportion of industrial energy consumption or rich diversity of climate zones is less meteorologically sensitive, which makes it challenging to implement LTLF. To address this issue, we propose a decomposition framework with industrial and temperature-sensitive components separated from the electric load and modeled, respectively. For the industrial component, a seldom-used macroeconomic variable, industrial gross products, is fed into modeling. For the temperature-sensitive component, temperature and its variants are input for modeling via variable selection. With serial and parallel modeling schemes, we further propose two decomposition models under this framework, where one, denoted as DeSerial, models two components serially, while the other, denoted as DeParallel, models them in a parallel way. Compared with representative models, the proposed two decomposition models produce more accurate monthly and annual load forecasts for the selected region, whose system load is less sensitive to temperature but more to the industries.

Mathematical modeling of ozone decomposition processes in wastewater treatment: A lumped kinetic approach with initial ozone demand

Ozone-based processes involve complex reaction networks and exhibit matrix-dependent decomposition patterns when dosed to wastewater effluents. Existing models for ozone decomposition are either too complex or not sufficiently descriptive of the regime manifested during initial ozone demand and decay. Models incorporating detailed reactions have shown limited applicability due to the high number of initial states and state variables involved. This study introduces a new mathematical model that balances accuracy and complexity in describing the main phases of ozone decomposition in secondary wastewater effluents. Specifically, a simplified model is proposed based on lumped variables, including initial ozone demand, two classes of organic matter (slow and fast reacting), radical forming (and ozone-decomposing) compounds, radical scavengers, and a hypothetical target contaminant. Results revealed that the model can predict with reasonable accuracy both the rapid initial ozone demand (occurring at very short timescales <10 s) and the slower ozone decomposition processes (occurring at longer timescales >30 s). A global sensitivity analysis was also conducted to further model refinement and simplification, which led to the elimination of three reactions connected to the generation and consumption of radicals (threshold |RXY| ≥ 0.05). Satisfactorily low residual values (RMSE≤7·10−6 M) were observed in all cases. Finally, the adequacy of the model was further tested against an independent set of ozone decomposition experiments obtained from the literature, confirming its suitability in describing ozone decomposition in secondary wastewater effluents with initial ozone demand.

Study on the Mechanism of Agricultural Greenhouse Gas Emission Reduction Under Macro Emission Reduction Measures

Clarifying the impact of macro emission reduction measures on the mechanism of agricultural greenhouse gas emission reduction is of great significance in promoting climate change governance and the construction of a carbon emission reduction policy system. This paper explores the mechanism of important macro emission reduction measures based on a multi-level progressive factor decomposition perspective and designs a coupled model of computable general equilibrium and structural path decomposition to identify the key emission reduction paths of major macro emission reduction measures and to decompose the drivers that promote emission reduction in each path. This study found that: (1) The emission reduction effect of the combination of carbon tax, carbon sink and carbon capture, utilization, and storage macro emission reduction measures is dominated by the indirect emission reduction triggered by the industrial chain, accounting for 95.67% of the total agricultural GHG emission reduction, and the emission reduction effect is gradually weakened with the increase in the production level. (2) The emission intensity effect and the industrial structure effect are the main drivers of the macro emission reduction portfolio measures to promote emission reduction, but there are differences in the roles of the different drivers on the various production levels and different emission reduction pathways. (3) Vegetables, fertilizers, the light industry, and other key industries are the main agricultural greenhouse gas emission reduction contributing industries, of which the emission reduction contribution from citizen consumption is the largest, and the emission reduction is mainly achieved by influencing the demand path of the vegetable industry and the light industry to the upstream high-energy-consuming or high-emission industries. Therefore, there is a need to fully utilize the mechanisms that drive emission reduction at different production levels and pathways by each key factor and to take targeted measures to promote synergistic emission reduction among industries. In the short term, focus on enhancing the role of the emission intensity effect, while in the medium and long term, pay much attention to the positive role of the industrial structure effect on agricultural greenhouse gas emissions.

Life prediction of on-board supercapacitor energy storage system based on gate recurrent unit neural network using sparse monitoring data

With the increasing use of supercapacitor in transportation and energy sectors, service life prediction becomes an important aspect to consider. As the aging process of onboard supercapacitors is closely related to practical working conditions, the actual service life may be inconsistent with the cycle life measured in the laboratory. However, the low-quality onboard monitoring data recording the historical working conditions is usually sparse and fragmented, making it difficult to extract valuable information. In our previous study, we successfully obtained the characteristic parameters from sparse and fragmented data, whereas those characteristic parameters change periodically and couldn't be used directly for life prediction. In this paper, we firstly extract the degradation trend term of supercapacitor by a composite sine and polynomial time series decomposition model from the characteristic parameters. Secondly, in order to make up for the lack of data, a GRU network is designed to generate more sample data which is in consistent with historical data evolution trends. The combination of input characteristic variables including the extracted historical characteristic capacitance C, temperature T and the time fitting sequences CtD are selected to improve the accuracy of GRU predictions. The predictive error of the characteristic capacitance C is 2.36 %. Finally, the life prediction of on-board supercapacitors based on actual working conditions is realized.

A Stacking Ensemble Learning Algorithm for Anomaly Early Detection of Liquid Levels in High-Pressure Heaters of Turbine Reheat Systems

In the steam turbine reheating system, maintaining the high-pressure heater liquid level is crucial to prevent equipment damage and safety accidents. High liquid levels may cause steam to enter the feed system, while low liquid levels may lead to overheating and damage to the heater tube bundle. In this study, the stacking ensemble learning model for combined data decomposition is proposed for high-precision liquid level prediction and early warning. Initially, outliers in the high-pressure heater liquid level data from the DSC system are processed. The Adaptive Local Iterative Filtering (ALIF) algorithm is used for the first data decomposition, followed by Empirical Mode Decomposition (EMD) for a second decomposition. Time-domain features such as summation, maximum value, skewness, and peak value are extracted to construct reconstructed data. Random Forest (RF) reduces the dimensionality of the reconstructed data. The prediction model combines Deep Belief Network (DBN), Deep Neural Networks (DNN), Light Gradient Boosting Machine (LGBM), and eXtreme Gradient Boosting (xGBoost) as base models, with Echo State Network (ESN) as the meta-learner. Finally, based on historical data, warning level values are set, and the system outputs warnings when the actual liquid level significantly deviates from the predicted value.

Word and morpheme frequency effects in naming Mandarin Chinese compounds: More than a replication

The question whether compound words are stored in our mental lexicon in a decomposed or full-listing way prompted Janssen and colleagues (2008) to investigate the representation of compounds using word and morpheme frequencies manipulations. Our study replicated their study using a new set of stimuli from a spoken corpus and incorporating EEG data for a more detailed investigation. In the current study, despite ERP analyses revealing no word frequency or morpheme frequency effects across conditions, behavioral outcomes indicated that Mandarin compounds are not sensitive to word frequency. Instead, the response times highlighted a morpheme frequency effect in naming Mandarin compounds, which contrasted with the findings of Janssen and colleagues. These findings challenge the full-listing model and instead support the decompositional model.