Optimized Ensemble Extreme Learning Machine for Classification of Electrical Insulators Conditions

Stefenon, Stéfano Frizzo; Grebogi, Rafael Bartnik; Freire, Roberto Zanetti; Nied, Ademir; Meyer, Luiz Henrique

doi:10.1109/tie.2019.2926044

Cited by 59 publications

(25 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thereafter, the components obtained in the previous step (IMFs and one residue component) are trained using extreme learning machines (ELM) [42], SVR [43], Gaussian process (GP) [44], and gradient boosting machines (GBM) [45]. These individual models are chosen due to the effects already observed for regression and time series forecasting tasks, as described in [46][47][48].…”

Section: Objective and Contributionmentioning

confidence: 99%

See 1 more Smart Citation

Electricity Price Forecasting Based on Self-Adaptive Decomposition and Heterogeneous Ensemble Learning

et al. 2020

Self Cite

View full text Add to dashboard Cite

Electricity price forecasting plays a vital role in the financial markets. This paper proposes a self-adaptive, decomposed, heterogeneous, and ensemble learning model for short-term electricity price forecasting one, two, and three-months-ahead in the Brazilian market. Exogenous variables, such as supply, lagged prices and demand are considered as inputs signals of the forecasting model. Firstly, the coyote optimization algorithm is adopted to tune the hyperparameters of complementary ensemble empirical mode decomposition in the pre-processing phase. Next, three machine learning models, including extreme learning machine, gradient boosting machine, and support vector regression models, as well as Gaussian process, are designed with the intent of handling the components obtained through the signal decomposition approach with focus on time series forecasting. The individual forecasting models are directly integrated in order to obtain the final forecasting prices one to three-months-ahead. In this case, a grid of forecasting models is obtained. The best forecasting model is the one that has better generalization out-of-sample. The empirical results show the efficiency of the proposed model. Additionally, it can achieve forecasting errors lower than 4.2% in terms of symmetric mean absolute percentage error. The ranking of importance of the variables, from the smallest to the largest is, lagged prices, demand, and supply. This paper provided useful insights for multi-step-ahead forecasting in the electrical market, once the proposed model can enhance forecasting accuracy and stability.

show abstract

Section: Objective and Contributionmentioning

confidence: 99%

“…In this approach, hidden nodes are randomly chosen and outputs are obtained analytically. Good generalization and fast learning speed are the main advantages of ELM [48]. The input weights and hidden biases are specified arbitrarily and then are fixed.…”

Section: Extreme Learning Machinementioning

confidence: 99%

Electricity Price Forecasting Based on Self-Adaptive Decomposition and Heterogeneous Ensemble Learning

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…Using a combination of extreme learning machine [5] and quantile regression, the research presented in [6] shows an evaluation of PV forecasting with high accuracy. The work is based on a linear programming optimisation model that presents an efficient and much faster result than traditional methods like backpropagation neural networks (NNs).…”

Section: Introductionmentioning

confidence: 99%

Hybrid deep learning for power generation forecasting in active solar trackers

Stefenon

Kasburg²,

Nied

et al. 2020

IET Generation, Transmission & Distribution

Self Cite

View full text Add to dashboard Cite

“…Machine learning algorithms such as extreme learning machine (ELM) [24], random forest (RF) [25], and logistic regression [26] have become an important research issue and have been widely applied in a variety of fields [27]. In this paper, we choose support vector regression (SVR) to establish temperature drift models for MEMS gyroscopes.…”

Section: Introductionmentioning

confidence: 99%

Dual-Mass MEMS Gyroscope Parallel Denoising and Temperature Compensation Processing Based on WLMP and CS-SVR

2020

View full text Add to dashboard Cite

For the sake of decreasing the effects of noise and temperature error on the measurement accuracy of micro-electro-mechanical system (MEMS) gyroscopes, a denoising and temperature drift compensation parallel model method based on wavelet transform and forward linear prediction (WFLP) and support vector regression based on the cuckoo search algorithm (CS-SVR) is proposed in this paper. First, variational mode decomposition (VMD) is proposed in this paper, which is aimed at dividing the output signal of the gyroscope into intrinsic mode functions (IMFs); then, the IMFs are classified into three features—drift, mixed, and pure noise features—by the sample entropy (SE) value. Second, a wavelet transform and forward linear prediction (WFLP) are combined to remove the noise from the mixed features. Meanwhile, the drift feature is compensated by support vector regression based on the cuckoo search algorithm (CS-SVR). Finally, through reconstruction, the final signal is obtained. Experimental results demonstrate that the VMD-SE-WFLP-CS-SVR method proposed in this paper can decrease noise and compensate the temperature error effectively (angular random walking value is optimized from 1.667°/√h to 0.0667°/√h and the bias stability is reduced from 30°/h to 4°/h). In terms of denoising, the performance of the WFLP algorithm is superior to the wavelet threshold and FLP, as it combines their advantages; furthermore, in terms of temperature compensation, the proposed CS-SVR algorithm uses the cuckoo search algorithm to find the optimal parameters of SVR, improving the accuracy of the model.

show abstract

Optimized Ensemble Extreme Learning Machine for Classification of Electrical Insulators Conditions

Cited by 59 publications

References 32 publications

Electricity Price Forecasting Based on Self-Adaptive Decomposition and Heterogeneous Ensemble Learning

Electricity Price Forecasting Based on Self-Adaptive Decomposition and Heterogeneous Ensemble Learning

Hybrid deep learning for power generation forecasting in active solar trackers

Dual-Mass MEMS Gyroscope Parallel Denoising and Temperature Compensation Processing Based on WLMP and CS-SVR

Contact Info

Product

Resources

About