Multivariate chaotic time series prediction based on improved extreme learning machine

Zhang, Ruiquan; Xu, Mingming; Han, Min; Li, Hongxing

doi:10.23919/chicc.2017.8027985

Cited by 5 publications

(3 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…That is to say, the closer the prediction points are, the greater the impact of training samples on prediction results will be. 19 Therefore, there is necessity conducting implicit weighting on the input samples of KELM model on the time scale. Concretely, weighting coefficients are used to weigh the prediction error variables of the model, and then the objective function 15 can be rewritten as follows:…”

Section: Kelm and Implicit Weightingmentioning

confidence: 99%

“…Han and Wang 19 indicate that the KELM model ignores that different prediction window sample points have a different influence upon the prediction performance and these sample points exhibit heteroscedasticity which can be harmful to the fault prognostics. Furthermore, Han and Wang 19 point out that it's necessary to assign different weights to the sample points in time scale to improve the prediction accuracy. This significant standpoint shows us a new road to the higher prediction accuracy.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Rolling Bearing Fault Trend Prediction Based on Composite Weighted KELM

Yu¹,

Li²,

Tian³

et al. 2018

IJAV

View full text Add to dashboard Cite

The abilities of different degradation feature types to characterize rolling bearing fault trend are distinctive. And even the characteristic ability of the same degradation feature can change at various times. Thus these feature samples possess heteroscedasticity. However, traditional kernel extreme learning machine (KELM) model assumes that different input samples' effects on the predicted value are equal, which results in low prediction precision and low computing efficiency. To solve this problem, a novel composite weighted KELM (CWKELM) prediction model, which is fused with explicit weighting and implicit weighting, is proposed. In both feature type scale and sample time scale, the feature samples and the prediction model are weighted according to the prediction error. An adaptive mutation particle swarm optimization (AMPSO) algorithm is applied in optimizing the penalty factor and the kernel parameter in the model. Taking various entropy features as the input samples, the proposed model is adopted to conduct one-step and multi-step prediction for rolling bearing fault trend. Experimental results show that this prediction model has higher prediction accuracy and computing efficiency compared with the traditional KELM model.

show abstract

Section: Kelm and Implicit Weightingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Rolling Bearing Fault Trend Prediction Based on Composite Weighted KELM

Yu¹,

Li²,

Tian³

et al. 2018

IJAV

View full text Add to dashboard Cite

show abstract

“…In this paper, in order to know the spreading direction of the fire, it is necessary to determine whether the ultra-short-term wind speed and direction of the well field in different terrains and different time scales have chaotic properties. If a nonlinear system produces an initial value at a certain moment, and all other values produced after this initial value are related to this initial value, showing sensitivity to the initial value, then the nonlinear system has chaotic properties, and the time series with chaotic properties is called chaotic time series [21]. Chaotic time series Sample Number is a kind of complex time series, although it seems to be chaotic and irregular, but it has infinite levels of self-similar structure, which is essentially deterministic time series with short-term predictability [22].…”

mentioning

confidence: 99%

Method for predicting ultra-short-term wind speed and direction at emergency well site

Pan,

Liang,

2024

Preprint

View full text Add to dashboard Cite

In order to protect the safety of emergency rescue personnel and improve the speed of well control emergency rescue, it is necessary to know the ultra-short-term wind speed and direction time series of the well field in the next 15 minutes in advance. In this paper, the 0-1 test method is used to calculate the chaotic properties of the ultra-short-term wind speed and direction in the well field, and then the Hurst index is used to calculate the predictability of the ultra-short-term wind speed and direction in the well field. The Seagull Optimization Algorithm (SOA) is used to optimize the four hyper-parameters of the Echo State Network (ESN) and to train the network output weight matrix, so that the ultra-short-term wind speed and direction in the well field can be predicted from chaotic time series of the three kinds of terrain of the well field at different time scales. Among the three terrains, the MSE of the ultra-short-term wind speed in the well field valley and the ultra-short-term wind direction in the well field mountain are the smallest, which are 0.0215 and 1.787, respectively. The experimental results show that the SOA-ESN method has a good prediction effect on the chaotic time series of the ultra-short-term wind speed and direction, and can provide a new technical support for speeding up the well control emergency rescue.

show abstract

Rank-Revealing Orthogonal Decomposition in Extreme Learning Machine Design

Kabziński

2018

Artificial Neural Networks and Machine Learning – ICANN 2018

View full text Add to dashboard Cite

Multivariate chaotic time series prediction based on improved extreme learning machine

Cited by 5 publications

References 16 publications

Rolling Bearing Fault Trend Prediction Based on Composite Weighted KELM

Rolling Bearing Fault Trend Prediction Based on Composite Weighted KELM

Method for predicting ultra-short-term wind speed and direction at emergency well site

Rank-Revealing Orthogonal Decomposition in Extreme Learning Machine Design

Contact Info

Product

Resources

About