Forecasting Current Velocity and Profile in a Strait Water using Warped Gaussian Process

2021 ◽  
pp. 1-35
Author(s):  
Kai Wei ◽  
Xiang Liao ◽  
Shunquan Qin
Keyword(s):  
Gaussian Process ◽  
Kernel Function ◽  
Current Velocity ◽  
Time Window ◽  
Offshore Structures ◽  
Kernel Functions ◽  
Ocean Current ◽  
Time Interval ◽  
Short Term ◽  
Gp Model

Abstract Ocean current forecast is vital for developing tidal energy and construction of offshore structures in the strait waters. This paper developed a short-term ocean current forecasting approach using the warped Gaussian process (WGP), which consists of the measured data preprocessing, kernel function selection, and data forecasting using WGP. A preprocessing using the wavelet thresholding method was proposed to enhance the quality of the measured raw data. The theory of WGP and the commonly used kernel functions were briefly introduced. The sliding time window and one-step ahead strategies were employed to increase the accuracy of predictions. Observations collected during an ocean current measurement campaign executed in a strait water on the coast of the East China Sea were used as an example dataset. The current velocity and profile were forecasted and validated using the example dataset as an illustration of the framework of the developed approach. The effects of window length, kernel function, and time interval on the WGP forecasting efficiency and precision were investigated. The forecasting performance of the developed WGP model was discussed by comparing it with the standard GP model. The current profile with a 95% confidence interval was also predicted by the developed WGP model at a certain point. The validation shows that the developed model is efficient in the short-term ocean current forecast.

A Comparison of Numerical Optimizers in Developing High Dimensional Surrogate Models

2019 ◽  
Author(s):  
Yanwen Xu ◽  
Pingfeng Wang
Keyword(s):  
Gaussian Process ◽  
Process Model ◽  
Likelihood Function ◽  
Surrogate Models ◽  
Kernel Functions ◽  
Function Optimization ◽  
High Dimensional ◽  
Model Parameters ◽  
Gradient Based ◽  
Gp Model

Abstract The Gaussian Process (GP) model has become one of the most popular methods to develop computationally efficient surrogate models in many engineering design applications, including simulation-based design optimization and uncertainty analysis. When more observations are used for high dimensional problems, estimating the best model parameters of Gaussian Process model is still an essential yet challenging task due to considerable computation cost. One of the most commonly used methods to estimate model parameters is Maximum Likelihood Estimation (MLE). A common bottleneck arising in MLE is computing a log determinant and inverse over a large positive definite matrix. In this paper, a comparison of five commonly used gradient based and non-gradient based optimizers including Sequential Quadratic Programming (SQP), Quasi-Newton method, Interior Point method, Trust Region method and Pattern Line Search for likelihood function optimization of high dimension GP surrogate modeling problem is conducted. The comparison has been focused on the accuracy of estimation, the efficiency of computation and robustness of the method for different types of Kernel functions.

A comparative study of Gaussian process regression with other three machine learning approaches in the performance prediction of centrifugal pump

2021 ◽  
pp. 095440622110505
Author(s):  
Xutao Zhao ◽  
Desheng Zhang ◽  
Renhui Zhang ◽  
Bin Xu
Keyword(s):  
Machine Learning ◽  
Gaussian Process ◽  
Centrifugal Pump ◽  
Kernel Function ◽  
Performance Prediction ◽  
Optimization Design ◽  
Gaussian Process Regression ◽  
Kernel Functions ◽  
Performance Indices ◽  
Pump Performance

Accurate prediction of performance indices using impeller parameters is of great importance for the initial and optimal design of centrifugal pump. In this study, a kernel-based non-parametric machine learning method named with Gaussian process regression (GPR) was proposed, with the purpose of predicting the performance of centrifugal pump with less effort based on available impeller parameters. Nine impeller parameters were defined as model inputs, and the pump performance indices, that is, the head and efficiency, were determined as model outputs. The applicability of three widely used nonlinear kernel functions of GPR including squared exponential (SE), rational quadratic (RQ) and Matern5/2 was investigated, and it was found by comparing with the experimental data that the SE kernel function is more suitable to capture the relationship between impeller parameters and performance indices because of the highest R square and the lowest values of max absolute relative error (MARE), mean absolute proportional error (MAPE), and root mean square error (RMSE). In addition, the results predicted by GPR with SE kernel function were compared with the results given by other three machine learning models. The comparison shows that the GPR with SE kernel function is more accurate and robust than other models in centrifugal pump performance prediction, and its prediction errors and uncertainties are both acceptable in terms of engineering applications. The GPR method is less costly in the performance prediction of centrifugal pump with sufficient accuracy, which can be further used to effectively assist the design and manufacture of centrifugal pump and to speed up the optimization design process of impeller coupled with stochastic optimization methods.

Study of Sequential Minimal Optimization Algorithm Type and Kernel Function Selection for Short-Term Load Forecasting

2013 ◽  
Vol 329 ◽  
pp. 472-477
Author(s):  
Zhe Zhao ◽  
Xiao Yu Li
Keyword(s):  
Kernel Function ◽  
Thermal Power ◽  
Load Forecasting ◽  
Nonlinear Function ◽  
Optimal Solution ◽  
Kernel Functions ◽  
Support Vector ◽  
Sequential Minimal Optimization ◽  
Short Term ◽  
Short Term Load Forecasting

Short-term load forecasting is important for power system operation,including preparing plans for generation and supply, arranging the generator to set start or stop, coordinating thermal power units and hydropower units. Support vector machines have advantage in approximating any nonlinear function with arbitrary precision and modeling by studying history data. Based on SVM, this paper selects the sequential minimal optimization (SMO) algorithm to compute load forecasting, because SMO can avoid iterative, so as to short the running time. If we select different kernel functions and the SMO type in the computing process, we will receive different result. Though the analysis of results,the paper obtains the optimal solution in different accuracy or time requirements for short-term load forecasting. By a power plant data, respectively, it discusses from the weekly load forecasts and daily load forecast to play an empirical analysis. It concludes that the selection of ɛ-SVR type and the linear form kernel function is ideal for short-term load forecasting in a not strictly time limits. Otherwise, it will select others in different terms.

Frequency Hopping Prediction Based on Multi-Kernel SVM

2013 ◽  
Vol 336-338 ◽  
pp. 2256-2260
Author(s):  
Yu Kai Yao ◽  
Yong Qing Yu ◽  
Yang Liu ◽  
Jin Jin Wang ◽  
Xiao Yun Chen
Keyword(s):  
Phase Space ◽  
Kernel Function ◽  
Frequency Hopping ◽  
Phase Space Reconstruction ◽  
Kernel Functions ◽  
Chaotic Attractors ◽  
Short Term ◽  
Regression Problem ◽  
Frequency Hopping Sequences ◽  
Kernel Svm

The chaotic frequency hopping sequences possesses short-term predictability. Via the phase space reconstruction approach, we can get chaos attractors, and the problem of series prediction can be transformed into the regression problem of the chaotic attractors. This paper uses SVR method to deal with the prediction of frequency hopping sequences. After analyzing the characteristics of existing kernel functions, we produce a new multi-kernel function, which is used for the prediction of frequency hopping sequences. Experiments show the fine performances of our methods.

Time Window from Visual Images to Visual Short-Term Memory: Consolidation or Integration?

2004 ◽  
Vol 51 (1) ◽  
pp. 45-51 ◽  
Author(s):  
Yuhong Jiang
Keyword(s):  
Short Term Memory ◽  
Time Window ◽  
Short Interval ◽  
Visual Images ◽  
Short Term ◽  
Temporal Window ◽  
Term Memory ◽  
Visual Short Term Memory ◽  
Probe Task ◽  
Long Time

Abstract. When two dot arrays are briefly presented, separated by a short interval of time, visual short-term memory of the first array is disrupted if the interval between arrays is shorter than 1300-1500 ms ( Brockmole, Wang, & Irwin, 2002 ). Here we investigated whether such a time window was triggered by the necessity to integrate arrays. Using a probe task we removed the need for integration but retained the requirement to represent the images. We found that a long time window was needed for performance to reach asymptote even when integration across images was not required. Furthermore, such window was lengthened if subjects had to remember the locations of the second array, but not if they only conducted a visual search among it. We suggest that a temporal window is required for consolidation of the first array, which is vulnerable to disruption by subsequent images that also need to be memorized.

scholarly journals Time Interval to Biochemical Failure as a Surrogate End Point in Locally Advanced Prostate Cancer: Analysis of Randomized Trial NRG/RTOG 9202

2019 ◽  
Vol 37 (3) ◽  
pp. 213-221 ◽  
Author(s):  
James J. Dignam ◽  
Daniel A. Hamstra ◽  
Herbert Lepor ◽  
David Grignon ◽  
Harmar Brereton ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Randomized Trial ◽  
Locally Advanced ◽  
Biochemical Failure ◽  
Mediating Effect ◽  
Time Interval ◽  
Short Term ◽  
End Points ◽  
End Point

Background In prostate cancer, end points that reliably portend prognosis and treatment benefit (surrogate end points) can accelerate therapy development. Although surrogate end point candidates have been evaluated in the context of radiotherapy and short-term androgen deprivation (AD), potential surrogates under long-term (24 month) AD, a proven therapy in high-risk localized disease, have not been investigated. Materials and Methods In the NRG/RTOG 9202 randomized trial (N = 1,520) of short-term AD (4 months) versus long-term AD (LTAD; 28 months), the time interval free of biochemical failure (IBF) was evaluated in relation to clinical end points of prostate cancer–specific survival (PCSS) and overall survival (OS). Survival modeling and landmark analysis methods were applied to evaluate LTAD benefit on IBF and clinical end points, association between IBF and clinical end points, and the mediating effect of IBF on LTAD clinical end point benefits. Results LTAD was superior to short-term AD for both biochemical failure (BF) and the clinical end points. Men remaining free of BF for 3 years had relative risk reductions of 39% for OS and 73% for PCSS. Accounting for 3-year IBF status reduced the LTAD OS benefit from 12% (hazard ratio [HR], 0.88; 95% CI, 0.79 to 0.98) to 6% (HR, 0.94; 95% CI, 0.83 to 1.07). For PCSS, the LTAD benefit was reduced from 30% (HR, 0.70; 95% CI, 0.52 to 0.82) to 6% (HR, 0.94; 95% CI, 0.72 to 1.22). Among men with BF, by 3 years, 50% of subsequent deaths were attributed to prostate cancer, compared with 19% among men free of BF through 3 years. Conclusion The IBF satisfied surrogacy criteria and identified the benefit of LTAD on disease-specific survival and OS. The IBF may serve as a valid end point in clinical trials and may also aid in risk monitoring after initial treatment.

Inter-Repetition Interval in Short-Term Motor Memory

1975 ◽  
Vol 40 (2) ◽  
pp. 535-538 ◽  
Author(s):  
Philip H. Marshall ◽  
Susan L. Wyatt ◽  
Shirley A. Moore ◽  
Stephen E. Sigman
Keyword(s):  
Short Interval ◽  
Motor Memory ◽  
Time Interval ◽  
Short Term ◽  
Verbal Responses ◽  
Motor Movement ◽  
Repetition Interval ◽  
Improved Accuracy

An investigation was conducted to ascertain the influence of the duration of the time interval between successive repetitions of a discrete motor movement in a short-term motor memory paradigm. With one repetition a long interval increased error relative to a short interval. The opposite was true for seven repetitions; a long interval improved accuracy. The results were discussed in terms of the “trace shrinkage” hypothesis and compared with those from similar studies using verbal responses.

scholarly journals A Novel Methodology for Hydrocarbon Depth Prediction in Seabed Logging: Gaussian Process-Based Inverse Modeling of Electromagnetic Data

2021 ◽  
Vol 11 (4) ◽  
pp. 1492
Author(s):  
Hanita Daud ◽  
Muhammad Naeim Mohd Aris ◽  
Khairul Arifin Mohd Noh ◽  
Sarat Chandra Dass
Keyword(s):  
Gaussian Process ◽  
Inverse Modeling ◽  
Mean Squared Error ◽  
Computational Effort ◽  
Percentage Error ◽  
Depth Prediction ◽  
Receiver System ◽  
Realistic Representation ◽  
And Performance ◽  
Gp Model

Seabed logging (SBL) is an application of electromagnetic (EM) waves for detecting potential marine hydrocarbon-saturated reservoirs reliant on a source–receiver system. One of the concerns in modeling and inversion of the EM data is associated with the need for realistic representation of complex geo-electrical models. Concurrently, the corresponding algorithms of forward modeling should be robustly efficient with low computational effort for repeated use of the inversion. This work proposes a new inversion methodology which consists of two frameworks, namely Gaussian process (GP), which allows a greater flexibility in modeling a variety of EM responses, and gradient descent (GD) for finding the best minimizer (i.e., hydrocarbon depth). Computer simulation technology (CST), which uses finite element (FE), was exploited to generate prior EM responses for the GP to evaluate EM profiles at “untried” depths. Then, GD was used to minimize the mean squared error (MSE) where GP acts as its forward model. Acquiring EM responses using mesh-based algorithms is a time-consuming task. Thus, this work compared the time taken by the CST and GP in evaluating the EM profiles. For the accuracy and performance, the GP model was compared with EM responses modeled by the FE, and percentage error between the estimate and “untried” computer input was calculated. The results indicate that GP-based inverse modeling can efficiently predict the hydrocarbon depth in the SBL.

Procedural Complications During Early Versus Late Endovascular Treatment in Acute Stroke

Stroke ◽  
2021 ◽  
Author(s):  
Errikos Maslias ◽  
Stefania Nannoni ◽  
Federico Ricciardi ◽  
Bruno Bartolini ◽  
Davide Strambo ◽  
...  
Keyword(s):  
Endovascular Treatment ◽  
Acute Stroke ◽  
Time Window ◽  
National Institutes Of Health ◽  
Late Time ◽  
Short Term ◽  
Term Outcome ◽  
Stroke Registry ◽  
Short Term Outcome ◽  
Procedural Complications

Background and Purpose: Endovascular treatment (EVT) in acute ischemic stroke is effective in the late time window in selected patients. However, the frequency and clinical impact of procedural complications in the early versus late time window has received little attention. Methods: We retrospectively studied all acute ischemic strokes from 2015 to 2019 receiving EVT in the Acute Stroke Registry and Analysis of Lausanne. We compared the procedural EVT complications in the early (<6 hours) versus late (6–24 hours) window and correlated them with short-term clinical outcome. Results: Among 695 acute ischemic strokes receiving EVT (of which 202 were in the late window), 113 (16.3%) had at least one procedural complication. The frequency of each single, and for overall procedural complications was similar for early versus late EVT (16.2% versus 16.3%, P adj =0.90). Procedural complications lead to a significantly less favorable short-term outcome, reflected by the absence of National Institutes of Health Stroke Scale improvement in late EVT (delta-National Institutes of Health Stroke Scale-24 hours, −2.5 versus 2, P adj =0.01). Conclusions: In this retrospective analysis of consecutive EVT, the frequency of procedural complications was similar for early and late EVT patients but very short-term outcome seemed less favorable in late EVT patients with complications.

Sign in / Sign up

Export Citation Format

Share Document