scholarly journals Day-Ahead Wind Speed Forecasting Using Relevance Vector Machine

2014 ◽  
Vol 2014 ◽  
pp. 1-6 ◽  
Author(s):  
Guoqiang Sun ◽  
Yue Chen ◽  
Zhinong Wei ◽  
Xiaolu Li ◽  
Kwok W. Cheung
Keyword(s):  
Wind Speed ◽  
Kernel Function ◽  
Back Propagation ◽  
Relevance Vector Machine ◽  
Back Propagation Neural Network ◽  
Gaussian Kernel ◽  
Polynomial Kernel ◽  
Support Vector ◽  
Stable Operation ◽  
Wind Speed Forecasting

With the development of wind power technology, the security of the power system, power quality, and stable operation will meet new challenges. So, in this paper, we propose a recently developed machine learning technique, relevance vector machine (RVM), for day-ahead wind speed forecasting. We combine Gaussian kernel function and polynomial kernel function to get mixed kernel for RVM. Then, RVM is compared with back propagation neural network (BP) and support vector machine (SVM) for wind speed forecasting in four seasons in precision and velocity; the forecast results demonstrate that the proposed method is reasonable and effective.

ARRHYTHMIA DISEASE DIAGNOSIS USING NEURAL NETWORK, SVM, AND GENETIC ALGORITHM-OPTIMIZED k-MEANS CLUSTERING

2011 ◽  
Vol 11 (04) ◽  
pp. 897-915 ◽  
Author(s):  
ROSHAN JOY MARTIS ◽  
CHANDAN CHAKRABORTY
Keyword(s):  
Neural Network ◽  
Genetic Algorithm ◽  
Normal Sinus Rhythm ◽  
Optimization Technique ◽  
Back Propagation ◽  
Principal Component ◽  
Disease Diagnosis ◽  
Back Propagation Neural Network ◽  
Polynomial Kernel ◽  
Support Vector

This work aims at presenting a methodology for electrocardiogram (ECG)-based arrhythmia disease detection using genetic algorithm (GA)-optimized k-means clustering. The open-source ECG data from MIT-BIH arrhythmia database and MIT-BIH normal sinus rhythm database are subjected to a sequence of steps including segmentation using R-point detection, extraction of features using principal component analysis (PCA), and pattern classification. Here, the classical classifiers viz., k-means clustering, error back propagation neural network (EBPNN), and support vector machine (SVM) have been initially attempted and subsequently m-fold (m = 3) cross validation is used to reduce the bias during training of the classifier. The average classification accuracy is computed as the average over all the three folds. It is observed that EBPNN and SVM with different order polynomial kernel provide significant accuracies in comparison with k-means one. In fact, the parameters (centroids) of k-means algorithm are locally optimized by minimizing its objective function. In order to overcome this limitation, a global optimization technique viz., GA is suggested here and implemented to find more robust parameters of k-means clustering. Finally, it is shown that GA-optimized k-means algorithm enhances its accuracy to those of other classifiers. The results are discussed and compared. It is concluded that the GA-optimized k-means algorithm is an alternate approach for classification whose accuracy will be near to that of supervised (viz., EBPNN and SVM) classifiers.

scholarly journals Short-Term Wind Speed Forecasting Based on Signal Decomposing Algorithm and Hybrid Linear/Nonlinear Models

Energies ◽  
2018 ◽  
Vol 11 (11) ◽  
pp. 2976 ◽  
Author(s):  
Qinkai Han ◽  
Hao Wu ◽  
Tao Hu ◽  
Fulei Chu
Keyword(s):  
Neural Network ◽  
Wind Speed ◽  
Nonlinear Models ◽  
Moving Average ◽  
Back Propagation ◽  
Percentage Error ◽  
Support Vector ◽  
Intrinsic Mode Functions ◽  
Short Term ◽  
Wind Speed Forecasting

Accurate wind speed forecasting is a significant factor in grid load management and system operation. The aim of this study is to propose a framework for more precise short-term wind speed forecasting based on empirical mode decomposition (EMD) and hybrid linear/nonlinear models. Original wind speed series is decomposed into a finite number of intrinsic mode functions (IMFs) and residuals by using the EMD. Several popular linear and nonlinear models, including autoregressive integrated moving average (ARIMA), support vector machine (SVM), random forest (RF), artificial neural network with back propagation (BP), extreme learning machines (ELM) and convolutional neural network (CNN), are utilized to study IMFs and residuals, respectively. An ensemble forecast for the original wind speed series is then obtained. Various experiments were conducted on real wind speed series at four wind sites in China. The performance and robustness of various hybrid linear/nonlinear models at two time intervals (10 min and 1 h) are compared comprehensively. It is shown that the EMD based hybrid linear/nonlinear models have better accuracy and more robust performance than the single models with/without EMD. Among the five hybrid models, EMD-ARIMA-RF has the best accuracy on the whole for 10 min data, and the mean absolute percentage error (MAPE) is less than 0.04. However, for the 1 h data, no model can always perform well on the four datasets, and the MAPE is around 0.15.

Research on the Fouling Prediction Based on Hybrid Kernel Function Relevance Vector Machine

2011 ◽  
Vol 204-210 ◽  
pp. 31-35
Author(s):  
Ling Fang Sun ◽  
Hong Gang Xie ◽  
Li Hong Qiao
Keyword(s):  
Kernel Function ◽  
Relevance Vector Machine ◽  
Economic Benefits ◽  
Training Data ◽  
Gaussian Kernel ◽  
Polynomial Kernel ◽  
Simulation Research ◽  
Dynamic Simulator ◽  
Hybrid Kernel Function ◽  
Hybrid Kernel

The research on the fouling prediction of heat exchanger is significantly to improve operational efficiency and economic benefits of the plants. Based on the relevance vector machine with Gaussian kernel function, polynomial kernel function and hybrid kernel function, simulation research on the fouling prediction was introduced. We construct a six-inputs and one-output network model according to the fouling monitor principle and parameters with MATLAB, all training data came from the Automatic Dynamic Simulator of Fouling and input the network after normalized processing and reclassification. Simulations show that the root mean square error of fouling prediction with hybrid kernel function is less than simple kernel function, and has the better prediction precision.

scholarly journals Electrical Characteristics Estimation of Photovoltaic Modules via Cuckoo Search—Relevant Vector Machine Probabilistic Model

2021 ◽  
Vol 9 ◽  
Author(s):  
Jianmin Ban ◽  
Xinyu Pan ◽  
Ziqiang Bi ◽  
Minming Gu
Keyword(s):  
Neural Network ◽  
Back Propagation ◽  
Cuckoo Search ◽  
Relevance Vector Machine ◽  
Back Propagation Neural Network ◽  
Computational Effort ◽  
Accurate Estimation ◽  
Support Vector ◽  
Electrical Characteristics ◽  
Characteristics Estimation

This work presents an optimized probabilistic modeling methodology that facilitates the modeling of photovoltaic (PV) modules with measured data over a range of environmental conditions. The method applies cuckoo search to optimize kernel parameters, followed by electrical characteristics estimation via relevance vector machine. Unlike analytical modeling techniques, the proposed cuckoo search-relevance vector machine (CS-RVM) takes advantages of no required knowledge of internal PV parameters, more accurate estimation capability and less computational effort. A comparative study has been done among the electrical characteristics predicted by back-propagation neural network (BPNN), radial basis function neural network (RBFNN), support vector machine (SVM), Villalva's model, relevance vector machine (RVM), and the CS-RVM. Experimental results show that the proposed CS-RVM provides the best prediction in most scenarios.

Study on Application of Multi-Kernel Relevance Vector Machine in Data Validation of Thermal System in Power Plant

2014 ◽  
Vol 672-674 ◽  
pp. 1421-1424
Author(s):  
Yue Zhao ◽  
Feng Qi Si ◽  
Zhi Gao Xu
Keyword(s):  
Power Plant ◽  
Kernel Function ◽  
Relevance Vector Machine ◽  
Test Method ◽  
Gaussian Kernel ◽  
Polynomial Kernel ◽  
Thermal System ◽  
Data Validation ◽  
Data Regression ◽  
Hybrid Kernel

A new method for data validation of thermal process in power plant was proposed based on multi-kernel relevance vector machine (MKRVM). Hybrid kernel function combining Gaussian kernel and polynomial kernel function was applied to relevance vector machine (RVM). After optimizing kernel parameters with firefly algorithm, nonlinear data regression model was built for thermal system. Then we used measuring test method to detect data and reconstruct data with model prediction. This method can overcome multiple correlation and nonlinearity among variables of thermal system. It has good robustness against various types of noise and higher accuracy for prediction compared with SVM and RVM. The results of case analysis for thermal system in a 600 MW unit show this method can detect and reconstruct data effectively.

scholarly journals Identification of Pilots’ Fatigue Status Based on Electrocardiogram Signals

Sensors ◽  
2021 ◽  
Vol 21 (9) ◽  
pp. 3003
Author(s):  
Ting Pan ◽  
Haibo Wang ◽  
Haiqing Si ◽  
Yao Li ◽  
Lei Shang
Keyword(s):  
Back Propagation ◽  
Principal Component ◽  
Back Propagation Neural Network ◽  
Practical Significance ◽  
Support Vector ◽  
State Identification ◽  
Feature Parameter ◽  
The Time Domain ◽  
Pilot Fatigue ◽  
Identification Model

Fatigue is an important factor affecting modern flight safety. It can easily lead to a decline in pilots’ operational ability, misjudgments, and flight illusions. Moreover, it can even trigger serious flight accidents. In this paper, a wearable wireless physiological device was used to obtain pilots’ electrocardiogram (ECG) data in a simulated flight experiment, and 1440 effective samples were determined. The Friedman test was adopted to select the characteristic indexes that reflect the fatigue state of the pilot from the time domain, frequency domain, and non-linear characteristics of the effective samples. Furthermore, the variation rules of the characteristic indexes were analyzed. Principal component analysis (PCA) was utilized to extract the features of the selected feature indexes, and the feature parameter set representing the fatigue state of the pilot was established. For the study on pilots’ fatigue state identification, the feature parameter set was used as the input of the learning vector quantization (LVQ) algorithm to train the pilots’ fatigue state identification model. Results show that the recognition accuracy of the LVQ model reached 81.94%, which is 12.84% and 9.02% higher than that of traditional back propagation neural network (BPNN) and support vector machine (SVM) model, respectively. The identification model based on the LVQ established in this paper is suitable for identifying pilots’ fatigue states. This is of great practical significance to reduce flight accidents caused by pilot fatigue, thus providing a theoretical foundation for pilot fatigue risk management and the development of intelligent aircraft autopilot systems.

Sign in / Sign up

Export Citation Format

Share Document