Self-adaptive Extreme Learning Machine Research Articles

Solar Radiation Prediction in Adrar, Algeria: A Case Study of Hybrid Extreme Machine-Based Techniques

This study delves into the application of hybrid extreme machine-based techniques for solar radiation prediction in Adrar, Algeria. The models under evaluation include the Extreme Learning Machine (ELM), Weighted Extreme Learning Machine (WELM), and Self-Adaptive Extreme Learning Machine (SA-ELM), with a comparative analysis based on various performance metrics. The results show that SA-ELM achieves the highest accuracy with an R2 of 0.97, outperforming ELM and WELM by 4.6% and 15.4% respectively in terms of R2. SA-ELM also has the lowest MPE, RMSE and RRMSE values, indicating a higher accuracy in predicting global radiation. Furthermore, comparison with previously employed prediction techniques solidifies SA-ELM’s superiority, evident in its 0.275 RMSE.The study explores different input combinations for predicting global radiation in the study region, concluding that incorporating all relevant inputs yields optimal performance, although reduced input scenarios can still provide practical accuracy when data availability is limited. These results highlight the effectiveness of the SA-ELM model in accurately predicting global radiation, which is expected to have significant implications for renewable energy applications in the region. However, further testing and evaluation of the models in different regions and under different weather conditions is recommended to improve the generalizability and robustness of the results.

Study of water resources parameters using artificial intelligence techniques and learning algorithms: a survey

Qualitative analysis of water resources is one of the most widely used topics in water resources research today. Researchers use various analysis methods of water parameters to achieve the desired goals in this field. This research uses artificial intelligence (AI), learning machine (LM), data mining, and mathematical techniques to simulate water behavior and estimate its parametric changes. The proposed model used in this study was a Self-adaptive Extreme learning machine (SAELM) to estimate hydrogeological parameters of the Meghan wetland located in Markazi province in Iran. In addition, SAELM simulation results were compared to Least square support vector machine (LSSVM), Multiple linear regression (MLR), and Adaptive Neuro-fuzzy inference system (ANFIS) models. The simulated parameters were Electrical Conductivity (EC), Total Dissolved Solids (TDS), Groundwater Level (GWL), and salinity. This information was related to sampling for 175 months in the study area. Finally, after simulation operation, four models were introduced as superior models. Mentioned exceptional models were SAELM in GWL modeling, SAELM in modeling the EC, MLR in salinity simulation, and LSSVM in the simulation of TDS parameters. Moreover, by five approaches, the models' performance was evaluated. Suggested strategies were performance evaluation by statistical indicators, Wilson score method uncertainty analysis (WSMUA), response & correlation plots, discrepancy ratio charts, and distribution error diagrams. Based on statistical indicators, the SAELMGWL model was the most accurate model with RMSE, MAPE, and R2 indices equal to 0.1496, 0.0043, and 0.9933, respectively. The ANFIS model had the worst results in simulation.

A non-tuned machine learning method to simulate ice-seabed interaction process in clay

Exploitation of oil and gas in the Arctic area is expected to expand in the coming years. These hydrocarbons are transferred through subsea pipelines from offshore to onshore; however, the marine pipelines are threatened by traveling icebergs where the seabed may be gouged by the moving masses in warmer months. Subsea trenching and backfilling are usually utilized to bury the subsea pipelines for physical protection against the ice scouring. Regarding the stress-based design methods, deformations and forces are generally the controlling design factors for the subsea assets. In this study, the subgouge clay displacements and the reaction forces were simulated using a non-tuned self-adaptive machine learning (ML) entitled “self-adaptive extreme learning machine” (SAELM). Initially, fifteen SAELM models were defined by means of the parameters affecting the ice-scoured features. Subsequently, 70% and 30% of the constructed dataset were respectively applied to train and test the ML models. After that, the optimum number of hidden layer neurons and the best activation function were selected for the SAELM network. By conducting a comprehensive sensitivity analysis, the premium SAELM models and the most influencing input parameters in estimation of the subgouge clay characteristics were introduced. Regarding the performed analyses, the horizontal load factor and the gouge depth ratio were identified as the most influential parameters to model the reaction forces, whereas the soil depth had a significant impact for simulation of the ice-induced clay deformations. Finally, a set of SAELM-based equations were presented to estimate the subgouge clay parameters.

Estimation of water level fluctuations in groundwater through a hybrid learning machine

In the present study, a new hybrid artificial intelligence-based model is presented to simulate monthly time-series data relevant to groundwater quantity collected from an observation well located in Kermanshah, Iran. To define the hybrid AI model, the differential evolutionary is applied to optimize the Extreme Learning Machine (ELM) so that the Self Adaptive ELM (SAELM) is developed and the wavelet transform is employed to decompose the input variables of the model (WSAELM). The autocorrelation function is utilized for detecting effective lags of the time-series data. Moreover, 70% of the observed data is used to train the AI techniques and the rest (i.e. 30%) is applied in test mode. Then, using these influencing lags, different models are defined for the SAELM and WSAELM models. After that, superior models for the simulation of GWL are selected by evaluating different input combinations. For instance, the values of the Nash-Sutcliffe efficiency coefficient, variance accounted for (VAF) and correlation coefficient (R) for the superior WSAELM model are obtained to be 0.973, 97.450 and 0.988, respectively. The prediction uncertainty of the developed SAELM model for GWL is quantified with a value of ±0.081. Finally, a formula is developed for computing GWL. The obtained results show that the applied hybrid machine learning model can be utilized to estimate the groundwater quantity parameter (i.e. groundwater level) in the Karnachi well and the possible applicability of the WSAELM technique to other areas.

RGB-D based human action recognition using evolutionary self-adaptive extreme learning machine with knowledge-based control parameters

RGB-D based human action recognition using evolutionary self-adaptive extreme learning machine with knowledge-based control parameters

Self-adaptive extreme learning machine-based prediction of roller length of hydraulic jump on rough bed

ABSTRACT In this study, the roller length of hydraulic jumps occurring on rough beds is modeled using the Self-Adaptive Extreme Learning Machine (SAELM) method. For this purpose, the parameters influencing the roller length are specified and four different SAELM models are developed based on them. A superior model is also established by analyzing the modeling results. For the superior model, the statistical values of the Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and correlation coefficient are calculated to be 1.720, 6.369 and 0.969, respectively. Also, the results of the SAELM superior model are compared with the Multilayer Perceptron Neural Network (MLPNN) and Support Vector Machine (SVM) methods. The analysis of the SVM, MLPNN and SVM models results reveals the effectiveness of the SAELM model. In this study, the uncertainty analysis of the SAELM, MLPNN and SVM models is also performed and the prediction error interval of 95% for the SAELM model is obtained which varies from −0.112 to +0.134.

Estimation of total dissolved solids, electrical conductivity, salinity and groundwater levels using novel learning machines

In this study, the groundwater parameters including electrical conductivity (EC), salinity, total dissolved solids (TDS) and groundwater level (GWL) for a 15-year time series (from 2002 to 2017) of the Mighan Plain, located in Markazi Province, Iran, were simulated using a hybrid meta heuristic artificial intelligence (AI) model called “wavelet self-adaptive extreme learning machine” (WSAELM) for the first time. Initially, the detection of the most significant lags of the time-series data was conducted using the autocorrelation function (ACF) and the partial autocorrelation function (PACF) analyses. By using these lags, four WSAELM models were defined and then the superior models in simulating the TDS, EC, salinity and GWL were introduced. The values of the determination coefficient (R2), Variance Accounted for (VAF) and the Nash–Sutcliffe efficiency coefficient (NSC) for the superior model simulating salinity were computed to be 0.991, 98.124 and 0.980, respectively. Also, approximately 44% of the TDS values modeled by the best model had an error less than 10%, while roughly a third of the TDS values estimated by the model had an error more than 15%. The findings indicated that the proposed hybrid model underestimated the GWL parameter, while it performed in an overestimate way for other parameters. The results of the uncertainty analysis showed the low width of uncertainty for GWL (WUB = ± 0.798), TDS (WUB = ± 5.035), EC (WUB = ± 6.425) and salinity (WUB = ± 66.650).

A NEW NON‐TUNED SELF‐ADAPTIVE MACHINE‐LEARNING APPROACH FOR SIMULATING THE DISCHARGE COEFFICIENT OF LABYRINTH WEIRS

AbstractIn this study, the labyrinth weir discharge coefficient was simulated using the self‐adaptive extreme learning machine (SAELM) artificial intelligence model in two cases: normal orientation labyrinth weirs (NLWs) and inverted orientation labyrinth weirs (ILWs). First, the most optimized neuron of the hidden layer was computed. The number of hidden layer neurons was calculated as 30. Also, by analysing the results of different activation functions, it was concluded that the sigmoid activation function has higher accuracy than the others. Next, the superior model was identified by conducting a sensitivity analysis. The model approximated the discharge coefficient of labyrinth weirs with reasonable accuracy. For example, the R2, scatter index and Nash–Sutcliffe efficiency coefficient for the best model were estimated as 0.966, 0.034 and 0.964, respectively. In addition, the ratio of the total head above the weir to the height of the weir crest (HT/P) and the ratio of length of apex geometry to width of a single cycle (A/w) were identified as the most effective parameters. Furthermore, the uncertainty analysis results indicated that the superior model had an overestimated performance. Then, a relationship was proposed in terms of all input variables for the superior model. © 2020 John Wiley & Sons, Ltd.

Forecasting of Groundwater Level Using Ensemble Hybrid Wavelet–Self-adaptive Extreme Learning Machine-Based Models

Generally, the estimation and prediction of groundwater levels (GWLs) are an important challenge in the field of water resources management. In this study, for the first time, the groundwater level (GWL) of the Sarab Qanbar well located in the south of Kermanshah, Iran, is approximated using the wavelet–SAELM model. An artificial intelligence method called “self-adaptive extreme learning machine (SAELM)” and the “wavelet transform” method were implemented to develop a predictive model. First, using the autocorrelation function (ACF) and the partial autocorrelation function (PACF), the effective lags in estimating the GWL were detected and evaluated. Then, eight distinctive SAELM and WA–SAELM models were developed using the input combinations selected by the ACF and the PACF as different lags. Later, the values of the observational well were normalized to estimate the GWL. Time series data were divided into two subsamples including 70% to train the artificial intelligence (AI) techniques and 30% to test the AI models. Next, the most optimized mother wavelet was chosen for the modeling. By evaluating the results of the SAELM and WA–SAELM models, it was concluded that compared to the SAELM models, the WA–SAELM models estimate values of the objective function with higher accuracy. Then, the superior model was introduced, which was very accurate in forecasting the GWL. In the test mode, for example, the correlation coefficient, mean absolute error and the Nash–Sutcliffe efficiency coefficient for the superior model were calculated as 0.995, 0.988 and 0.990, respectively. Furthermore, an uncertainty analysis was conducted for the predictive models, revealing that the superior model has an underestimate performance.

Simulation of discharge coefficient of side weirs placed on convergent canals using modern self-adaptive extreme learning machine

Side weirs are broadly used in irrigation channels, drainage systems and sewage disposal canals for controlling and adjusting the flow in main channels. In this study, a new artificial intelligence model entitled “self-adaptive extreme learning machine” (SAELM) is developed for simulating the discharge coefficient of side weirs located upon rectangular channels. Also, the Monte Carlo simulations are implemented for assessing the abilities of the numerical models. It should be noted that the k-fold cross-validation approach is used for validating the results obtained from the numerical models. Based on the parameters affecting the discharge coefficient, six artificial intelligence models are defined. The examination of the numerical models exhibits that such models simulate the discharge coefficient valued with acceptable accuracy. For instance, mean absolute error and root mean square error for the superior model are computed 0.022 and 0.027, respectively. The best SAELM model predicts the discharge coefficient values in terms of Froude number (Fd), ratio of the side weir height to the downstream depth (w/hd), ratio of the channel width at downstream to the downstream depth (bd/hd) and ratio of the side weir length to the downstream depth (L/hd). Based on the sensitivity analysis results, the Froude number of the side weir downstream is identified as the most influencing input parameter. Lastly, a matrix is presented to estimate the discharge coefficient of side weirs on convergent channels.

Estimation of scour depth around submerged weirs using self-adaptive extreme learning machine

Abstract In this study, the equilibrium scour depth downstream of the weir (ds-a), the maximum scour depth downstream of the weir (ds-max), the equilibrium scour depth upstream of the weir (dus-a) and the maximum scour depth upstream of the weir (dus-max) were simulated around the submerged weirs using the self-adaptive extreme learning machine (SAELM) model. In other words, the SAELM was utilized for the simulation of the scour depths around submerged weirs for the first time. In addition, Monte Carlo simulations (MCSs) were used to increase the accuracy of the artificial intelligence model. The results of modeling were validated using k-fold cross validation. At first, all effective parameters on the scour depth were determined and five distinct SAELM models were defined. Then, the optimal activation function of the SAELM model was obtained. By analyzing the results of modeling, the best models were identified to estimate ds-a/ht, ds-max/ht, dus-a/ht, and dus-max/ht, and the ratio of the average inflow velocity to the critical velocity (U0/Uc) was determined as the most effective input parameter. In the following, the results of superior models were compared with the artificial neural network (ANN) and support vector machine (SVM). The results showed that SAELM models were more accurate. The uncertainty analysis was performed for these models, some of them were overestimated and others were underestimated. In addition, some equations were presented for equilibrium models for calculation of scour depth around the submerged weirs, which are used by environmental and hydraulic engineers without previous knowledge about the artificial intelligence models. Finally, a partial derivative sensitivity analysis (PDSA) was performed for all input parameters of the superior models.

Evaluation of side orifices shape factor using the novel approach self-adaptive extreme learning machine

Generally, side orifices are used to regulate and measure the flow within open channels. In this paper, discharge coefficient of side orifices was modeled using the novel method “Self-Adaptive Extreme Learning Machine” for the first time. In current study, ability of the numerical model was evaluated using Monte Carlo simulations. Additionally, the validation of the soft computing model was done using the k-fold cross validation approach. The value of k was considered equal to 5. Then, effective parameters on discharge coefficient were identified. Additionally, the number of the hidden layer neurons was selected equal to 30. Also, the sigmoid activation function was chosen as the most optimum activation function. Using input parameters, five different self-adaptive extreme learning machine (SAELM) models were developed. Additionally, the shape factor of the side orifice was studied. The results of numerical models showed that models with shape factor simulated discharge coefficient with higher accuracy. After that, the superior model (SAELM 1) was introduced and R, RMSE, and MAPE for this model were computed 0.995, 0.004, and 0.092. Moreover, the MAE for the superior model was calculated 0.0004. The superior model predicted the discharge coefficient using all input parameters. The results of the sensitivity analysis showed that the ratio of flow depth in main channel to side orifice length Ym/D was identified as the most effective input parameter.

A Novel Approach for Prediction of Monthly Ground Water Level Using a Hybrid Wavelet and Non-Tuned Self-Adaptive Machine Learning Model

In recent decades, due to groundwater withdrawal in the Kabodarahang region, Iran, Hamadan, hazardous events such as sinkholes, droughts, water scarcity, etc., have occurred. This study models groundwater level (GWL) of the Kabodarahang region using two novel techniques including Self-Adaptive Extreme Learning Machine (SAELM) and Wavelet-Self-Adaptive Extreme Learning Machine (WA-SAELM). Using the stepwise selection as different lags along with different input combinations, ten different SAELM and WA-SAELM models were developed. First, the best activation function is chosen for numerical models. After that, GWL values were normalized to equalize the values and enhance speed and accuracy of modeling. Then, an optimized mother wavelet is selected in order to simulate GWLs. Next, the best model was introduced as the superior model in which values of the correlation coefficient (R), Root Mean Squared Error (RMSE) and Nash-Sutcliffe efficiency coefficient (NSC) were obtained 0.969, 0.358 and 0.939, respectively. In addition, the results of the superior model are compared with classical neural network models such as Artificial Neural Network (ANN), Wavelet-Artificial Neural Network (WA-ANN), Support Vector Machine (SVM) and Wavelet-Support Vector Machine (WA-SVM). Among all models, WA-SAELM approximated GWLs with higher accuracy. Furthermore, based on the results obtained from an uncertainty analysis, the superior model was identified as a model with an underestimated performance. Additionally, an explicit and practical matrix was proposed for computing GWLs. Finally, the matrix was validated for another piezometer.

Prediction of scour depth around bridge piers using self-adaptive extreme learning machine

Accurate prediction of pier scour can lead to economic design of bridge piers and prevent catastrophic incidents. This paper presents the application of self-adaptive evolutionary extreme learning machine (SAELM) to develop a new model for the prediction of local scour around bridge piers using 476 field pier scour measurements with four shapes of piers: sharp, round, cylindrical, and square. The model network parameters are optimized using the differential evolution algorithm. The best SAELM model calculates the scour depth as a function of pier dimensions and the sediment mean diameter. The developed SAELM model had the lowest error indicators when compared to regression-based prediction models for root mean square error (RMSE) (0.15, 0.65, respectively) and mean absolute relative error (MARE) (0.50, 2.0, respectively). The SAELM model was found to perform better than artificial neural networks or support vector machines on the same dataset. Parametric analysis showed that the new model predictions are influenced by pier dimensions and bed-sediment size and produce similar trends of variations of scour-hole depth as reported in literature and previous experimental measurements. The prediction uncertainty of the developed SAELM model is quantified and compared with existing regression-based models and found to be the least, ±0.03 compared with ±0.10 for other models.

Self-adaptive Extreme Learning Machine Optimized by Rough Set Theory and Affinity Propagation Clustering

Recently, a simple and efficient learning algorithm for single hidden layer feedforward networks (SLFNs) called extreme learning machine (ELM) has been developed by G.-B. Huang et al. One key strength of ELM algorithm is that there is only one parameter, the number of hidden nodes, to be determined while it has the significantly low computational time required for training new classifiers and good generalization performance. However, there is no effective method for finding the proper and universal number of hidden nodes. In order to address this problem, we propose a self-adaptive extreme learning machine (SELM) algorithm. SELM algorithm determines self-adaptively the number of hidden nodes and constructs Gaussian function as activation functions of hidden nodes. And in this algorithm, rough set theory acts as the pre-treatment cell to eliminate the redundant attributes of data sets. Then, affinity propagation clustering (AP Clustering) is used to self-adaptively determine the number of hidden nodes, while the centers and widths of AP clustering are utilized to construct a Gaussian function in the hidden layer of SLFNs. Empirical study of SELM algorithm on several commonly used classification benchmark problems shows that the proposed algorithm can find the proper number of hidden nodes and construct compact network classifiers, comparing with traditional ELM algorithm.

Improved probabilistic neural networks with self-adaptive strategies for transformer fault diagnosis problem

Probabilistic neural network has successfully solved all kinds of engineering problems in various fields since it is proposed. In probabilistic neural network, Spread has great influence on its performance, and probabilistic neural network will generate bad prediction results if it is improperly selected. It is difficult to select the optimal manually. In this article, a variant of probabilistic neural network with self-adaptive strategy, called self-adaptive probabilistic neural network, is proposed. In self-adaptive probabilistic neural network, Spread can be self-adaptively adjusted and selected and then the best selected Spread is used to guide the self-adaptive probabilistic neural network train and test. In addition, two simplified strategies are incorporated into the proposed self-adaptive probabilistic neural network with the aim of further improving its performance and then two versions of simplified self-adaptive probabilistic neural network (simplified self-adaptive probabilistic neural networks 1 and 2) are proposed. The variants of self-adaptive probabilistic neural networks are further applied to solve the transformer fault diagnosis problem. By comparing them with basic probabilistic neural network, and the traditional back propagation, extreme learning machine, general regression neural network, and self-adaptive extreme learning machine, the results have experimentally proven that self-adaptive probabilistic neural networks have a more accurate prediction and better generalization performance when addressing the transformer fault diagnosis problem.

Self-adaptive extreme learning machine

In order to overcome the disadvantage of the traditional algorithm for SLFN (single-hidden layer feedforward neural network), an improved algorithm for SLFN, called extreme learning machine (ELM), is proposed by Huang et al. However, ELM is sensitive to the neuron number in hidden layer and its selection is a difficult-to-solve problem. In this paper, a self-adaptive mechanism is introduced into the ELM. Herein, a new variant of ELM, called self-adaptive extreme learning machine (SaELM), is proposed. SaELM is a self-adaptive learning algorithm that can always select the best neuron number in hidden layer to form the neural networks. There is no need to adjust any parameters in the training process. In order to prove the performance of the SaELM, it is used to solve the Italian wine and iris classification problems. Through the comparisons between SaELM and the traditional back propagation, basic ELM and general regression neural network, the results have proven that SaELM has a faster learning speed and better generalization performance when solving the classification problem.

Age estimation based on children's voice: a fuzzy-based decision fusion strategy.

Automatic estimation of a speaker's age is a challenging research topic in the area of speech analysis. In this paper, a novel approach to estimate a speaker's age is presented. The method features a “divide and conquer” strategy wherein the speech data are divided into six groups based on the vowel classes. There are two reasons behind this strategy. First, reduction in the complicated distribution of the processing data improves the classifier's learning performance. Second, different vowel classes contain complementary information for age estimation. Mel-frequency cepstral coefficients are computed for each group and single layer feed-forward neural networks based on self-adaptive extreme learning machine are applied to the features to make a primary decision. Subsequently, fuzzy data fusion is employed to provide an overall decision by aggregating the classifier's outputs. The results are then compared with a number of state-of-the-art age estimation methods. Experiments conducted based on six age groups including children aged between 7 and 12 years revealed that fuzzy fusion of the classifier's outputs resulted in considerable improvement of up to 53.33% in age estimation accuracy. Moreover, the fuzzy fusion of decisions aggregated the complementary information of a speaker's age from various speech sources.

A novel self-adaptive extreme learning machine based on affinity propagation for radial basis function neural network

In this paper, a novel self-adaptive extreme learning machine (ELM) based on affinity propagation (AP) is proposed to optimize the radial basis function neural network (RBFNN). As is well known, the parameters of original ELM which developed by G.-B. Huang are randomly determined. However, that cannot objectively obtain a set of optimal parameters of RBFNN trained by ELM algorithm for different realistic datasets. The AP algorithm can automatically produce a set of clustering centers for the different datasets. According to the results of AP, we can, respectively, get the cluster number and the radius value of each cluster. In that case, the above cluster number and radius value can be used to initialize the number and widths of hidden layer neurons in RBFNN and that is also the parameters of coefficient matrix H of ELM. This may successfully avoid the subjectivity prior knowledge and randomness of training RBFNN. Experimental results show that the method proposed in this thesis has a more powerful generalization capability than conventional ELM for an RBFNN.