Parameter Estimation Model Research Articles

Constructing a heat source parameter estimation model for heat conduction finite element analysis using deep convolutional neural network

Heat conduction finite element analysis (FEA) is an important technique for estimating the temperature history of processes with local heat input, such as welding and additive manufacturing. This study proposes a framework to find appropriate heat source parameters without depending on the analyst’s skill. The heat source parameter estimation model consists of pre-trained deep convolutional and fully connected neural networks. The model determines appropriate heat source parameters such as the heat input, base metal shapes and temperature history. The model was constructed using a database created by heat conduction FEA. We demonstrated that heat source parameters were determined accurately for both known and unknown conditions.

ANALISIS SURVIVAL DENGAN COX PROPORTIONAL HAZARD PADA KASUS DEMAM TIFOID

A common problem found in survival data is the presence of censored data. The length of hospitalization of Typhoid fever patients until declared cured is one of example of this data. Here, we use Cox regression model to analysis this data. Partial likelihood is one of the methods of estimating parameters for Cox regression model. In many cases of censored data, two objects (patients) have the same length of hospitalization (ties). Therefore, to estimate the parameters of the model must use the right method. Here we used partial likelihood Breslow, Efron, and Exact methods. The study was motivated by how the three methods performed for Cox regression model. The data used for the implementation of these methods is length of hospitalization of Typhoid fever patients at Mekar Sari Hospital-Bekasi in 2020. Based on AIC criteria, we found that exact method is the best model (minimum AIC) for parameter estimation of Cox regression model. Referring to the Cox regression model by using a significance level of 10%, there are five predictor variables that affects the length of patient hospitalization. The five variables are age, vomiting, dirty tongue, hemoglobin, and leukocyte.Keywords: Typhoid fever, Cox regression, Breslow method, Efron method, exact method.MSC2020: 62N02, 62N03

Seismic Response Prediction of Open- and Closed-Loop Structural Control Systems Using a Multi-State-Dependent Parameter Estimation Approach

In this paper, a new modified multi-state-dependent parameter (MSDP) model estimation method is proposed with the aim of application for seismic response prediction of open- and closed-loop structural control systems. In the first phase, the recursive Kalman filtering is improved and an improved recursive fixed interval-smoothing algorithm is then proposed in the second phase. The numerical studies are conducted on a seismic-excited 3-story benchmark structure in three cases of open-loop structural system, open-loop structural control with tuned mass damper (TMD) as a passive device, and closed-loop structural control with an active tendon system (ATS). Using the proposed MSDP models, the time-histories of acceleration of the floors are predicted during earthquakes and the validation of the proposed models is evaluated with those given by the finite element method (FEM) of OpenSees software. Using three evaluation criteria, the superiority of the modified MSDP over the MSDP is found. The simulation results show that the proposed models are capable of following the seismic responses of the structure well and estimating them with high accuracy in further time steps. With high accuracy and lower computational cost, the surrogate models can be easily utilized in optimal design processes of the passive device for the open-loop structural control and the active control algorithm for the closed-loop structural control and structural health monitoring (SHM) applications.

Indoor occupancy estimation from carbon dioxide concentration using parameter estimation algorithms

The number of building occupants is an important indicator for predicting building energy consumption and developing control strategies for building automation. However, most occupancy estimation models were developed depending on the training steps where the true number of occupants is necessary. In order to calculate the occupant number independently, the newly-developed parameter estimation models were proposed, which are based on Maximum Likelihood (ML) approach and Bayesian analysis. A combination of multiple common measurements is used, including real-time CO2 concentration, energy consumption of facilities and make-up air system. The model starts by smoothing the raw CO2 concentration data using moving average, two-hour median as well as globally smooth. Then, ML and Bayesian analysis are used to establish the occupancy estimation models. The proposed models are evaluated in a commercial office which contains 36 occupants for validation. We find that the calculation errors could be reduced by using moving averaged data and globally smoothed data. The superiority of the parameter estimation models can be identified based on its lower calculation error and higher calculation accuracy compared to the previous established models. Practical Application Occupancy estimation models developed in this study are able to calculate occupant numbers independently and accurately in a non-intrusive way based on the indoor carbon dioxide concentration. This can provide input to a predictive building controller based on the application of occupancy estimation models. This could be applied to buildings across a district, informing demand-side management systems by employing occupancy behaviour and energy characteristics of individual buildings. This could allow both utility companies and building operators to simultaneously optimise their performance and benefit from this dedicated control strategy.

Deregulation of Interest Rate in Nigeria and Deposit Money Bank’s Performance (1996 – 2018)

This study looked at the relationship between interest rate deregulation and performance of Nigerian deposit money banks for the period 1996-2018. Interest rate deregulation was broken down into prime lending rate, maximum lending rate, 3-month deposit rate, and over 12-month deposit rate, with return on assets (ROA) serving as a proxy for deposit money bank performance. Dates for the variables listed above were obtained from the Central Bank of Nigeria Statistical Bulletin (2018 edition) and the World Bank data base. The data were assessed for stationarity with the Dickey-Fuller (D-F) test, long-run relationship with the Bound's co-integration test, and ARDL reliability with serial correlation, heteroscedasticity, and normality tests. The results of the tests revealed that all of the variables were integrated of order zero or one, and that the variables have a long-run relationship. As a result, the ARDL model for parameter estimation revealed that only prime lending rate was positively connected to bank ROA, whereas none of the explanatory factors were statistically significant. The researcher then claimed that there is no meaningful association between interest rate deregulation and the performance of Nigerian deposit money banks during the study period. As a result, deposit money institutions should attempt to mobilise appropriate savings from surplus units by offering deposit rates capable of enticing savers to grow their savings and increasing the availability of loanable funds.

Parameter sensitivity analysis of a physico-chemical lithium-ion battery model with combined discharge voltage and electrochemical impedance data

Physico-chemical battery models are capable of better simulation of lithium-ion cell behaviour due to their more physically descriptive modelling approach using many physical parameters. Accuracy of simulations highly depend on the correct identification of parameters. However, large parameter estimation uncertainties and model complexities increase the model parameterization difficulties. Therefore, pre-analysis of parameter sensitivities to model outputs can greatly help determining the best conditions and techniques for identifying many parameters with better accuracy. This work demonstrates a next step in improving the parameter identification accuracy by combined parameter sensitivity analysis of both discharge voltage and EIS data. A commercially available NMC-graphite based lithium-ion cell is modelled and 26 physical parameters are analysed for their sensitivity. For discharge voltage, capacity-related and non-capacity-related parameters clearly show different sensitivity patterns at different C-rates and depth of discharge regions. Moreover, non-dimensional EIS spectra states are introduced for correct identification of impedance-based parameter sensitivities and accurate parameter sensitivities are observed matching the frequency bands truly affected by the parameter kinetics. The ranked analysis of the parameter sensitivities clearly demonstrate the ability to identify many discharge voltage insensitive parameters using impedance measurements, which would greatly increase the number of identifiable parameters of a physico-chemical battery model.

A Neural Network Extension for Solving the Pareto/Negative Binomial Distribution Model

This research proposes a new estimation scheme to solve the estimation burden of the Pareto/NBD model and to release its power in out-of-sample prediction and then builds a neural network-based model (NNA-based model) for parameter estimation of the Pareto/NBD model, by designing a loss function to include the likelihood of the Pareto/NBD model and the mean absolute error. Following empirical analysis with a real dataset and simulation datasets, the main results are as follows. (1) The NNA-based model encounters a relatively skewed prediction in inactive prediction and focuses more on inactive prediction, while the heuristic method has a balanced prediction. (2) The Pareto/NBD model has better repeat purchase prediction compared to that under the mean absolute error, but the NNA-based model has a better fitting in the repeat purchase distribution. (3) The distribution of estimated parameters between these two models is quite different, but is able to realize the same prediction purpose. The NNA-based model also has greater application opportunities in the commercial environment because the learned estimator is applicable to different transaction timing patterns. These findings clarify that the NNA-based model solves the estimation burden and releases the predictive power of the Pareto/NBD model.

Nonlinear Reduced Models for State and Parameter Estimation

State estimation aims at approximately reconstructing the solution $u$ to a parametrized partial differential equation from $m$ linear measurements when the parameter vector $y$ is unknown. Fast numerical recovery methods have been proposed in Maday et al. [Internat. J. Numer. Methods Engrg., 102 (2015), pp. 933--965] based on reduced models which are linear spaces of moderate dimension $n$ that are tailored to approximate the solution manifold ${\cal M}$ where the solution sits. These methods can be viewed as deterministic counterparts to Bayesian estimation approaches and are proved to be optimal when the prior is expressed by approximability of the solution with respect to the reduced model [P. Binev et al., SIAM/ASA J. Uncertain. Quantif., 5 (2017), pp. 1--29]. However, they are inherently limited by their linear nature, which bounds from below their best possible performance by the Kolmogorov width $d_m({\cal M})$ of the solution manifold. In this paper, we propose to break this barrier by using simple nonlinear reduced models that consist of a finite union of linear spaces $V_k$, each having dimension at most $m$ and leading to different estimators $u_k^*$. A model selection mechanism based on minimizing the PDE residual over the parameter space is used to select from this collection the final estimator $u^*$. Our analysis shows that $u^*$ meets optimal recovery benchmarks that are inherent to the solution manifold and not tied to its Kolmogorov width. The residual minimization procedure is computationally simple in the relevant case of affine parameter dependence in the PDE. In addition, it results in an estimator $y^*$ for the unknown parameter vector. In this setting, we also discuss an alternating minimization (coordinate descent) algorithm for joint state and parameter estimation that potentially improves the quality of both estimators.

Coarse reduced model selection for nonlinear state estimation

State estimation is the task of approximately reconstructing a solu- tion u of a parametric partial differential equation when the parameter vector y is unknown and the only information is m linear measurements of u. Cohen et al. [arXiv:2009.02687, Nov. 2020] proposed a method to use a family of linear reduced spaces as a generalised nonlinear reduced model for state estimation. A computable surrogate distance is used to evaluate which linear estimate lies closest to a true solution of the pde problem. In this article we propose a strategy of coarse computation of the surrogate distance while maintaining a fine mesh reduced model, as the computational cost of the surrogate distance is large relative to the reduced modelling task. We demonstrate numerically that the error induced by the coarse distance is dominated by other approximation errors.

Strong-lensing source reconstruction with variationally optimized Gaussian processes

ABSTRACT Strong-lensing images provide a wealth of information both about the magnified source and about the dark matter distribution in the lens. Precision analyses of these images can be used to constrain the nature of dark matter. However, this requires high-fidelity image reconstructions and careful treatment of the uncertainties of both lens mass distribution and source light, which are typically difficult to quantify. In anticipation of future high-resolution data sets, in this work we leverage a range of recent developments in machine learning to develop a new Bayesian strong-lensing image analysis pipeline. Its highlights are (a) a fast, GPU-enabled, end-to-end differentiable strong-lensing image simulator; (b) a new, statistically principled source model based on a computationally highly efficient approximation to Gaussian processes that also takes into account pixellation; and (c) a scalable variational inference framework that enables simultaneously deriving posteriors for tens of thousands of lens and source parameters and optimizing hyperparameters via stochastic gradient descent. Besides efficient and accurate parameter estimation and lens model uncertainty quantification, the main aim of the pipeline is the generation of training data for targeted simulation-based inference of dark matter substructure, which we will exploit in a companion paper.

PARAMETER ESTIMATION OF NONLINEAR OUTPUT ERROR SYSTEM UNDER VARIATIONAL BAYESIAN METHOD BASED ON PROBABILISTIC GRAPHICAL MODEL

To estimate the parameters of the nonlinear output error system (nonlinear system), a variational Bayesian estimation method (VB method) is proposed based on the probabilistic graphical model (PGM). First, the related theories are introduced in this study such as the PGM and nonlinear systems. Then, the parameter estimation model of the nonlinear system is established. Finally, a VB method is proposed based on the PGM to estimate the parameters of the nonlinear system, which is tested and verified by numerical simulation experiments. It is found that the parameter estimation model of nonlinear system and the proposed method can estimate the parameters of relevant nonlinear system better, and the minimum error between the parameter estimated value and the actual value is only about 0.0001. It proves the feasibility of the VB method based on PGM in the identification of nonlinear systems. The results of this study provide an important reference for the control and identification of nonlinear systems.

Empirical Study on the Relationship between Agricultural Economic Structure Growth and Environmental Pollution Based on Time-Varying Parameter Vector Autoregressive Model.

In order to better demonstrate the relationship between agricultural economic structure growth and environmental pollution, an autoregressive model based on time-varying parameter vector was proposed. In the process of developing the research, this paper introduces the LDMI method, based on the time-varying parameter vector autoregression model, with the help of sampling formula calculation and other methods. Efforts were made to obtain credible conclusions. The experiment result shows that in this study, a total of 10,000 samples were taken. According to this value, 10000/116.15 = 86, which means that at least 86 unrelated samples can be obtained. Therefore, we can determine that each indicator mentioned in this paper has valid samples when it is introduced into the time-varying parameter vector autoregression (TVP-VAR) model for parameter estimation. After sampling detection image analysis and data calculation, the effect of energy structure, energy intensity industrial structure, and scale effect on the emission scale of environmental pollutants was obtained. It is proved that through the research of this paper, two main conclusions are finally obtained, and the influence of the five factors mentioned above is summarized.

WiSen: Zero-Knowledge Passive Human Tracking Using a Single WiFi Link

Some pioneer WiFi-based passive human tracking systems have achieved sub-meter level accuracy. However, their key limitation is that the location of WiFi devices must be known in advance. To this end, we propose WiSen, a novel system that estimates the location of a passive user in an indoor environment without any prior knowledge about the location of the WiFi devices. Specifically, we first exploit a signal power model for human-induced reflection signal extraction and multi-dimensional parameter estimation. Due to measurement noise and low-resolution parameter estimates, we further design a confidence-aware-based path pruning method that combines the distribution of path parameters from successive windows to select reliable human-induced reflection paths. Before that, we also adopt a novel data augmentation method to increase the number of available path parameters for better learning parameter distribution. These path parameters after path pruning are then employed to statistically estimate the relative location of the transmitter with respect to the receiver and finally localize the passive user. We implement WiSen on commodity WiFi devices and validate its performance in real indoor environments. The experimental results show that WiSen can realize the sub-meter level accuracy for passive human tracking and device localization with only a single WiFi link.

Parametric Software Effort Estimation Based on Optimizing Correction Factors and Multiple Linear Regression

<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Context:</i> Effort estimation is one of the essential phases that must be accurately predicted in the early stage of software project development. Currently, solving problems that affect the estimation accuracy of Use Case Points-based methods is still a challenge to be addressed. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Objective:</i> This paper proposes a parametric software effort estimation model based on Optimizing Correction Factors and Multiple Regression Models to minimize the estimation error and the influence of unsystematic noise, which has not been considered in previous studies. The proposed method takes advantage of the Least Squared Regression models and Multiple Linear Regression models on the Use Case Points-based elements. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Method:</i> We have conducted experimental research to evaluate the estimation accuracy of the proposed method and compare it with three previous related methods, i.e., 1) the baseline estimation method – Use Case Points, 2) Optimizing Correction Factors, and 3) Algorithmic Optimization Method. Experiments were performed on datasets (Dataset D1, Dataset D2, and Dataset D3). The estimation accuracy of the methods was analysed by applying various unbiased evaluation criteria and statistical tests. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Results:</i> The results proved that the proposed method outperformed the other methods in improving estimation accuracy. Statistically, the results proved to be significantly superior to the three compared methods based on all tested datasets. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Conclusion:</i> Based on our obtained results, the proposed method has a high estimation capability and is considered a helpful method for project managers during the estimation phase. The correction factors are considered in the estimation process.

Applying a Deep Learning Network in Continuous Physiological Parameter Estimation Based on Photoplethysmography Sensor Signals

In this paper, we propose a continuous physiological parameter estimation model based on a deep learning network for photoplethysmography (PPG) sensor signals. Signals of 8-s duration were incorporated into the proposed model in this study for frequent estimation of the systolic blood pressure (BP), diastolic BP, heart rate (HR), and mean arterial pressure of the human body; this facilitated early identification and monitoring of physiological conditions and thus reduced the risk of cardiovascular disease. The proposed model was designed using a convolutional neural network (CNN) and long short-term memory (LSTM) network. This model was trained and validated using the large-scale Multiparameter Intelligent Monitoring in Intensive Care database. The CNN was used to extract features from PPG signals automatically. This automatic extraction replaced the conventional manual feature extraction process. Features with time-series were then analyzed using the LSTM network to estimate physiological parameters. Subsequently, ten-fold cross-validation was conducted to reveal the mean absolute errors ± standard deviations of participants’ systolic BP, diastolic BP, HR, and mean arterial pressure to be 2.54 ± 3.88, 1.59 ± 2.45, 1.62 ± 2.55, and 1.59 ± 2.34 mmHg, respectively. These values meet the standards established by the Association for the Advancement of Medical Instrumentation and the British Hypertension Society. The proposed method facilitates the accurate, continuous monitoring of the BP and HR.

Distribution Feeder Parameter Estimation Without Synchronized Phasor Measurement by Using Radial Basis Function Neural Networks and Multi-Run Optimization Method

In practice, the performance of distribution feeder parameter estimation is limited by the measurement conditions in distribution networks. An accurate mathematical model that considers limited phasor measurements in distribution networks is necessary to estimate feeder parameters. This paper presents a set of modified parameter estimation models for unbalanced three-phase distribution feeders that only require the measurements of voltage amplitudes and power flows. To simplify the calculation process and improve the estimated results, a method combined with a radial basis function neural network (RBFNN) and multi-run optimization method (MRO), namely RBFNN-MRO, is proposed to calculate the parameters of distribution feeders. The relationship between the feeder parameters and the measurement data from the two terminals of the feeder can be mapped perfectly using the RBFNN. Furthermore, the random errors in the measurement device were eliminated using the proposed RBFNN-MRO algorithm. The RBFNN-MRO algorithm can limit the number of neurons in the hidden layer and substantially reduce the training time for each RBFNN. The feasibility of the proposed method was verified using four IEEE test systems. The proposed RBFNN-MRO and RBFNN methods were compared using the maximum absolute percentage error (MAPE) curves. The results reveal that the proposed RBFNN-MRO method has excellent potential for improving the accuracy of feeder parameter estimation even without synchronized phasor measurement.

Evaluation of Aircraft Auxiliary Power Unit Near-Field Acoustics for Condition Monitoring

This paper presents a comprehensive evaluation of the near-field acoustics of an aircraft auxiliary power unit (APU), based on experimental data acquired from an in-situ APU. The aim is to establish whether near-field acoustics can be implemented for online condition monitoring. The APU of Cranfield University’s demonstrator aircraft, a Boeing 737-400, has been instrumented to acquire acoustics (near-field and far-field) and vibration data in synchronization with aircraft state parameters under a wide range of operating conditions. The acquired data is first implemented to determine the efficacy of employing near-field / far-field microphones, and vibration sensors, to monitor the combustion noise and tonal frequency levels from the APU components. Subsequently, an evaluation of the broadband characteristics of the vibroacoustic data and its variations against APU states and performance parameters is conducted based on several categories of feature extraction techniques. The findings suggest that nearfield acoustics lacks the ability to capture the combustion noise process. In addition, the tonal frequencies are also lost due to the level of background noise, fluctuations in the APU speeds, and scattering effects. For the same reasons, the phase couplings occurring between the signals generated by the APU components cannot be detected using acoustic data. Nevertheless, the overall analysis substantiates that the near-field acoustic data can be used to predict the APU operating states and has the potential to be implemented for developing APU performance parameter estimation models to enable condition monitoring.

A Benchmark of Non-intrusive Parametric Audio Quality Estimation Models for Broadcasting Systems and Web-casting Applications

Due to the rising usage of various broadcasting systems and web-casting applications, a measurement of audio quality has become an essential task. This paper presents a benchmark of the parametric models for non-intrusive estimation of the audio quality perceived by the end user. The proposed solution is based on machine learning techniques for broadcasting systems and web-casting applications. The main goal of this study is to assess the performance of the non-intrusive parametric models as well as to evaluate a statistical significance of the performance differences between those models. The paper provides a comparison of several models based on the Support Vector Regression, Genetic Programming, Multigene Symbolic Regression, Neural Networks and Random Forest. The obtained results indicate that among the investigated models the most accurate, although not the fastest ones, are the model based on Random Forest (a broadcast scenario) and the SVR-based model (a web-cast scenario). These models represent promising candidates for non-intrusive parametric audio quality assessment in the context of broadcasting systems and web-casting applications.

Climate-sensitive diameter distribution models of larch plantations in north and northeast China

Tree diameter distribution models are important tools for forest management decision making. Climate change will affect tree growth and thus diameter distribution. However, there is lack of diameter distribution models describing the effects of climate. The paper developed a climate-sensitive diameter distribution model based on 193 sample plots of larch plantations in north and northeast China. Weibull distribution model was modified by a compatible simultaneous system and the percentile-based parameter recovery method with the inclusion of climate variables. The results showed that showed BIO10 (mean temperature of warmest quarter) andBIO14 (precipitation of driest month) are key climatic factors which shape the diameter distribution of Larch plantations. Compared with parameter estimation models without climate variables, the adjusted R2 was improved (5.03% for β, and 3.32 % for γ), RMSE (73.12% for β, and 2.89 % for γ)and TRE (92.16% for β, and 5.57% for γ) decreased. The models with climate factors are significantly different from that without climate factors by F test. The models with climate factors are improved and could be applied for diameter distribution projection under future climate change.

A Two‐Stage Bayesian Data‐Driven Method to Improve Model Prediction

AbstractBecause of the earth system complexity, groundwater/hydrology models are always built with structural errors, which may lead to systematic errors in model predictions. Bayesian data‐driven methods (DDMs) provide a feasible way to correct systematic model errors statistically. Generally, the physical and statistical model parameters, namely physical parameters and hyperparameters, are assumed to be independent and jointly calibrated. However, this assumption may be unreasonable and lead to over‐adjusted parameter estimation and biased model prediction. This study proposes a two‐stage DDM to calibrate physical parameters and hyperparameters separately, which does not make the independence assumption. Three case studies, including a groundwater solute transport analytical model, a three‐dimensional groundwater flow model, and a real‐world snowmelt runoff model, were used to evaluate the predictive performance of this two‐stage DDM. Based on the three case studies, we found that the independence assumption of physical parameters and hyperparameters could lead to the over‐fitting of parameter estimation and deviations in model predictions. Two‐stage DDM can constrain the systematic error model calibration; that is, physical parameters are first calibrated in the entire hyperparameter prior probability space, and then hyperparameters are calibrated in the posterior probability space of physical parameters obtained previously. As a result, compared with traditional joint calibration‐based DDM, two‐stage DDM can alleviate parameter over‐fitting and improve model predictive performance.