Incremental semi-supervised Extreme Learning Machine for Mixed data stream classification

Li, Qiude; Xiong, Qingyu; Ji, Shengfen; Yu, Yang; Wu, Chao; Gao, Min

doi:10.1016/j.eswa.2021.115591

Cited by 22 publications

(7 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to verify the effectiveness of the model, this paper sets up two comparative experiments: the first experiment compares the simulation results with the actual results, and the second experiment compares the results of LSTM-CA and ELM-CA. ELM-CA [61] trains the ELM model by collecting information about combustion and unburned areas and recording wind speed in the event of a fire. The probability is predicted by the action of elm and wind.…”

Section: The Results Of Lstm-ca and Elm-camentioning

confidence: 99%

Simulating Forest Fire Spread with Cellular Automation Driven by a LSTM Based Speed Model

Zhang

et al. 2022

Fire

View full text Add to dashboard Cite

The simulation of forest fire spread is a key problem for the management of fire, and Cellular Automata (CA) has been used to simulate the complex mechanism of the fire spread for a long time. The simulation of CA is driven by the rate of fire spread (ROS), which is hard to estimate, because some input parameters of the current ROS model cannot be provided with a high precision, so the CA approach has not been well applied yet in the forest fire management system to date. The forest fire spread simulation model LSTM-CA using CA with LSTM is proposed in this paper. Based on the interaction between wind and fire, S-LSTM is proposed, which takes full advantage of the time dependency of the ROS. The ROS estimated by the S-LSTM is satisfactory, even though the input parameters are not perfect. Fifteen kinds of ROS models with the same structure are trained for different cases of slope direction and wind direction, and the model with the closest case is selected to drive the transmission between the adjacent cells. In order to simulate the actual spread of forest fire, the LSTM-based models are trained based on the data captured, and three correction rules are added to the CA model. Finally, the prediction accuracy of forest fire spread is verified though the KAPPA coefficient, Hausdorff distance, and horizontal comparison experiments based on remote sensing images of wildfires. The LSTM-CA model has good practicality in simulating the spread of forest fires.

show abstract

Section: The Results Of Lstm-ca and Elm-camentioning

confidence: 99%

Simulating Forest Fire Spread with Cellular Automation Driven by a LSTM Based Speed Model

Zhang

et al. 2022

Fire

View full text Add to dashboard Cite

show abstract

“…By incorporating a substantial quantity of unlabeled samples lacking dominant variables into a limited number of labeled samples containing dominant variables, both the problem of multi-rate and the predictive accuracy of soft sensing models can be effectively addressed. This integration optimally utilizes the unlabeled samples, consequently resolving the challenge of multi-rate and enhancing the precision of prediction in soft sensing models [31,32]. In the selection of fusion weights, time is used as a variable, calculate its Pearson correlation coefficient, and the time correlation performance effectively reflects the degree of correlation between unlabeled and labeled samples.…”

Section: The Semi-supervised Learning Methods Based On Time Correlationmentioning

confidence: 99%

Research on soft sensor modeling method for complex chemical processes based on local semi-supervised selective ensemble learning

Liu,

Li,

Zhang

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

To address the adverse effects of nonlinearity and dynamic time-varying in complex chemical processes on the accuracy of the soft sensor model, a local-semi-supervised ensemble learning for soft sensor modeling (LS-SEL-LSTM) method is proposed in this article. Firstly, a hierarchical clustering method incorporating spatiotemporal criteria is proposed to reduce the influence of nonlinearity in global model prediction accuracy. The method considers the dynamic time-varying characteristics of temporal data and generates multiple local datasets. Then, to address the issue of multi-rate between auxiliary variables and dominant variables, a semi-supervised weight fusion mechanism based on temporal correlation is proposed, which effectively utilizes auxiliary variables to reconstruct local semi-supervised datasets and establishes local soft sensing models using LSTM. Concurrently, the parameters of the established model were optimized using the Flower Pollination Algorithm (FPA). Subsequently, a selective ensemble learning method based on sub-model prediction accuracy and an adaptive combination weight calculation method for sub-models were proposed to improve the prediction accuracy. Finally, the effectiveness of the proposed method was verified through the actual dataset of the sulfur recovery process. The results indicate that LS-SEL-LSTM performs well in handling complex chemical processes with nonlinear and dynamic time-varying characteristics.

show abstract

“…During this process, the model can continuously acquire new knowledge without retraining. Recently, incremental learning has been widely used in artificial immune algorithms, 32 incremental extreme learning machines, 33 incremental support vector machines (SVMs), 34 incremental artificial neural networks (NNs), and other fields. 35,36 In the above methods, inspired by the principle process of the vertebrate immune system and the need for self-organization and robustness of artificial intelligence systems, more and more researchers in the field of artificial intelligence focus on artificial immune algorithms.…”

Section: Introductionmentioning

confidence: 99%

A novel self-learning framework for fault identification of wind turbine drive bearings

Yuan

Liang

Wang

et al. 2023

Proceedings of the Institution of Mechanical Engineers, Part I:

View full text Add to dashboard Cite

The drive train is one of the core structures of a wind turbine, and its working condition seriously affects the performance quality. It is important to identify the fault pattern of the drive bearings in time to ensure the safety and reliability of the wind turbine. However, in traditional methods, offline modeling and online identification are often fragmented, and such a mechanism limits the adaptive updating of the model. To realize real-time updating and self-learning of the identification model, we proposed a novel self-learning framework for the intelligent fault identification of wind turbine drive bearings. First, a complete ensemble empirical mode decomposition with adaptive noise analysis–based quantification scheme for intrinsic mode function values is proposed. Then, based on the intrinsic mode function values, we offer an attention mechanism for fault feature identification and construct an initial fault pattern database using unsupervised clustering techniques. Second, abnormal data are identified by the proposed artificial immunity–based outlier detection algorithm to determine the type of immune response. Third, we design an automatic update strategy based on incremental learning to realize adaptive creation, deletion, and modification of fault patterns. The proposed intelligent framework is applied to the fault diagnosis of a real offshore wind turbine drive train, showing its advantages in intelligent fault identification and model updating.

show abstract

Incremental semi-supervised Extreme Learning Machine for Mixed data stream classification

Cited by 22 publications

References 45 publications

Simulating Forest Fire Spread with Cellular Automation Driven by a LSTM Based Speed Model

Simulating Forest Fire Spread with Cellular Automation Driven by a LSTM Based Speed Model

Research on soft sensor modeling method for complex chemical processes based on local semi-supervised selective ensemble learning

A novel self-learning framework for fault identification of wind turbine drive bearings

Contact Info

Product

Resources

About