Extreme Learning Machine Method Research Articles

Non-Tuned Machine Learning Approach for Predicting the Compressive Strength of High-Performance Concrete.

Compressive strength is considered as one of the most important parameters in concrete design. Time and cost can be reduced if the compressive strength of concrete is accurately estimated. In this paper, a new prediction model for compressive strength of high-performance concrete (HPC) was developed using a non-tuned machine learning technique, namely, a regularized extreme learning machine (RELM). The RELM prediction model was developed using a comprehensive dataset obtained from previously published studies. The input variables of the model include cement, blast furnace slag, fly ash, water, superplasticizer, coarse aggregate, fine aggregate, and age of specimens. k-fold cross-validation was used to assess the prediction reliability of the developed RELM model. The prediction results of the RELM model were evaluated using various error measures and compared with that of the standard extreme learning machine (ELM) and other methods presented in the literature. The findings of this research indicate that the compressive strength of HPC can be accurately estimated using the proposed RELM model.

A novel classification approach based on Extreme Learning Machine and Wavelet Neural Networks

In this paper, we present a novel classification approach based on Extreme Learning Machine (ELM) and Wavelet Neural Networks. We introduce two novel contributions. The first is Extreme Learning Machine based on Fast Wavelet Transform (ELM-FWT) algorithm aiming to solve the problem of matrix inversion which represents several limitations to ELM method. The second contribution is dedicated to solving another problem of machine learning algorithms, i.e. the number of neurons in the hidden nodes. It consists in allocating automatically and efficiently the number of neurons in the hidden layer. To demonstrate the effectiveness of our proposal, we realized numerous experiments and comparisons using 19 algorithms from Breast Cancer Wisconsin database which are considered as references in machine learning. Experimental results clearly demonstrate the stability and robustness of the proposed approach.

Event Identification by F-ELM Model for $\varphi$ -OTDR Fiber-Optic Distributed Disturbance Sensor

A novel event identification method, which combines the extreme learning machine (ELM) and fisher score feature selection method, is proposed to reduce the nuisance alarm rate (NAR) in fiber-optic distributed disturbance sensors based on phase-sensitive optical time-domain reflectometer (φ-OTDR). Through constructing 25.05 km long φ-OTDR experimental system and analyzing the selected features with the ELM, four kinds of real disturbance events, including watering, climbing, knocking and pressing, and a false disturbance event can be effectively identified. Experimental results show that the average identification rate of five disturbance events exceeds 95%, the identification time is below 0.1 s and the NAR is 4.67% through selecting 25 features. Compared with the ELM model without feature selection, the ELM model with feature selection by fisher method has several distinguished advantages of higher identification rate, shorter identification time, and lower NAR.

Prediction of the diet energy digestion using kernel extreme learning machine: A case study with Holstein dry cows

In order to effectively evaluate the diet nutritional and feeding value, it is essential to accurately predict the indicators of diet energy digestion of dairy cows, thus further optimizing feed formulation and improving feeding management. Traditional mathematical models used to predict the main indicators of diet energy digestion of dairy cows are usually based on linear regression (LR) method. However, as a typical parametric model, the LR-based method is limited by regression function assumption, and the inaccurate assumed function sometimes leads to the learned regression model biased from the ground-truth one. In this study, we propose a kernel extreme learning machine (KELM) technique to predict the indicators of Holstein dry cows' diet digestible energy (DE) and energy digestibility (ED). KELM is a typical non-parametric machine learning model, which does not require any specific assumptions about the regression function in advance. The learned modal by KELM can well fit the actual one only by learning the training sample data in most cases. To evaluate prediction accuracy effectively, we compared the KELM technique with traditional parametric prediction LR model and other commonly used non-parametric models such as radial basis function artificial neural network, support vector machine and standard extreme learning machine methods. The required sample data is obtained from actual feeding and digestion experiments. The prediction results indicate that the proposed KELM-based prediction technique is superior to other methods in most performance metrics for prediction of the DE and ED indicators of diet energy digestion of Holstein dry cows. In particular, it has higher prediction accuracy than traditional LR-based prediction method, and can predict DE and ED better under small samples conditions. The developed KELM-based prediction model may be utilized to provide key decision support information to animal nutrition experts, livestock farmers and feed suppliers. The presented model is also exploited as a potential tool for accurate evaluation of dairy cows diet feeding effect.

Hybrid Artificial Neural Network with Extreme Learning Machine Method and Cuckoo Search Algorithm to Predict Stock Prices

This thesis aims to predict the stock prices, using artificial neural network with extreme learning machine (ELM) method and cuckoo search algorithm (CSA). Stock is one type of investment that is in great demand in Indonesia. The portion ownership of stock is determined by how much investment is invested in the company. In this case, stock is an aggressive type of investment instrument, because stock prices can change over time. In this case, ELM is used to determine forecasting values, while CSA is applied to compile and optimize the values of weights and biases to be used in the forecasting process. After obtaining the best weights and biases, the validation test process is then carried out to determine the level of success of the training process. The data used is the daily data of the stock price of PT. Bank Mandiri (Persero) Tbk. the total is 291 data. Furthermore, the data is divided into 70% for the training process is as many as 199 data and 30% for the validation test as many as 87 data. Then compiled pattern of training and validation test patterns is 198 patterns and 82 patterns. Based on the implementation of the program, with several parameter obtained the result of MSE training is 0.001304353, with an MSE of validation test is 0.0031517704. Because the MSE value obtained is relatively small, this indicates that the ELM-CSA network is able to recognize data patterns and is able to predict well.

Fault Diagnosis of Fuel System Based on Improved Extreme Learning Machine

In this paper, extreme learning machine (ELM) method is used to classify the faults of fuel system. Although the learning speed of ELM is fast, its classification accuracy and generalization ability need to be improved. Bat Algorithm has a strong ability of global optimization. In order to make up for the deficiency of the ELM, this paper proposes a fault diagnosis model based on an improved bat algorithm to optimize the ELM. The experimental results show that the improved bat algorithm greatly improves the classification accuracy and generalization ability of the ELM, and verifies the validity of the proposed model.

Enhancing the Prediction Accuracy of Data-Driven Models for Monthly Streamflow in Urmia Lake Basin Based upon the Autoregressive Conditionally Heteroskedastic Time-Series Model

Hydrological modeling is one of the important subjects in managing water resources and the processes of predicting stochastic behavior. Developing Data-Driven Models (DDMs) to apply to hydrological modeling is a very complex issue because of the stochastic nature of the observed data, like seasonality, periodicities, anomalies, and lack of data. As streamflow is one of the most important components in the hydrological cycle, modeling and estimating streamflow is a crucial aspect. In this study, two models, namely, Optimally Pruned Extreme Learning Machine (OPELM) and Chi-Square Automatic Interaction Detector (CHAID) methods were used to model the deterministic parts of monthly streamflow equations, while Autoregressive Conditional Heteroskedasticity (ARCH) was used in modeling the stochastic parts of monthly streamflow equations. The state of art and innovation of this study is the integration of these models in order to create new hybrid models, ARCH-OPELM and ARCH-CHAID, and increasing the accuracy of models. The study draws on the monthly streamflow data of two different river stations, located in north-western Iran, including Dizaj and Tapik, which are on Nazluchai and Baranduzchai, gathered over 31 years from 1986 to 2016. To ascertain the conclusive accuracy, five evaluation metrics including Correlation Coefficient (R), Root Mean Square Error (RMSE), Nash–Sutcliffe Efficiency (NSE), Mean Absolute Error (MAE), the ratio of RMSE to the Standard Deviation (RSD), scatter plots, time-series plots, and Taylor diagrams were used. Standalone CHAID models have better results than OPELM methods considering sole models. In the case of hybrid models, ARCH-CHAID models in the validation stage performed better than ARCH-OPELM for Dizaj station (R = 0.96, RMSE = 1.289 m3/s, NSE = 0.92, MAE = 0.719 m3/s and RSD = 0.301) and for Tapik station (R = 0.94, RMSE = 2.662 m3/s, NSE = 0.86, MAE = 1.467 m3/s and RSD = 0.419). The results remarkably reveal that ARCH-CHAID models in both stations outperformed all other models. Finally, it is worth mentioning that the new hybrid “ARCH-DDM” models outperformed standalone models in predicting monthly streamflow.

Rapid discrimination of Salvia miltiorrhiza according to their geographical regions by laser induced breakdown spectroscopy (LIBS) and particle swarm optimization-kernel extreme learning machine (PSO-KELM)

Laser-induced breakdown spectroscopy (LIBS) coupled with particle swarm optimization-kernel extreme learning machine (PSO-KELM) method was developed for classification and identification of six types Salvia miltiorrhiza samples in different regions. The spectral data of 15 Salvia miltiorrhiza samples were collected by LIBS spectrometer. An unsupervised classification model based on principal components analysis (PCA) was employed first for the classification of Salvia miltiorrhiza in different regions. The results showed that only Salvia miltiorrhiza samples from Gansu and Sichuan Province can be easily distinguished, and the samples in other regions present a bigger challenge in classification based on PCA. A supervised classification model based on KELM was then developed for the classification of Salvia miltiorrhiza, and two methods of random forest (RF) and PSO were used as the variable selection method to eliminate useless information and improve classification ability of the KELM model. The results showed that PSO-KELM model has a better classification result with a classification accuracy of 94.87%. Comparing the results with that obtained by particle swarm optimization-least squares support vector machines (PSO-LSSVM) and PSO-RF model, the PSO-KELM model possess the best classification performance. The overall results demonstrate that LIBS technique combined with PSO-KELM method would be a promising method for classification and identification of Salvia miltiorrhiza samples in different regions.

A Fast Military Object Recognition using Extreme Learning Approach on CNN

Convolutional Neural Network (CNN) is an algorithm that can classify image data with very high accuracy but requires a long training time so that the required resources are quite large. One of the causes of the long training time is the existence of a backpropagation-based classification layer, which uses a slow gradient-based algorithm to perform learning, and all parameters on the network are determined iteratively. This paper proposes a combination of CNN and Extreme Learning Machine (ELM) to overcome these problems. Combination process is carried out using a convolution extraction layer on CNN, which then combines it with the classification layer using the ELM method. ELM method is Single Hidden Layer Feedforward Neural Networks (SLFNs) which was created to overcome traditional CNN’s weaknesses, especially in terms of training speed of feedforward neural networks. The combination of CNN and ELM is expected to produce a model that has a faster training time, so that its resource usage can be smaller, but maintaining the accuracy as much as standard CNN. In the experiment, the military object classification problem was implemented, and it achieves smaller resources as much as 400 MB on GPU comparing to standard CNN.

Study on optimization of rake head density of suction hopper dredger based on bat algorithm and extreme learning machine

The dredging output of suction dredger mainly comes from the suction density of the rake head. Accurate prediction of suction density is of great significance to improve the dredging output of suction dredger. In order to overcome the shortcomings of low accuracy and poor real-time performance of the current inhalation density prediction methods, a bat algorithm is proposed to optimize the inhalation density prediction method of extreme learning machine. The bat algorithms for optimizing extreme learning machines prediction model is constructed based on the measured construction data of “Xinhaifeng” Yangtze Estuary, and compared with other prediction models. Finally, the bat algorithms for optimizing extreme learning machines model is used to build the output simulator of inhalation density. Compared with the actual construction, the selection of control parameters is analyzed when the output of inhalation density is the best. Experients show that bat algorithms for optimizing extreme learning machines prediction has high accuracy and good stability, and can provide scientific and effective reference for yield prediction and construction guidance.

Correction: Discrimination of nasopharyngeal carcinoma serum using laser-induced breakdown spectroscopy combined with an extreme learning machine and random forest method

Correction for ‘Discrimination of nasopharyngeal carcinoma serum using laser-induced breakdown spectroscopy combined with an extreme learning machine and random forest method’ by Yanwu Chu et al., J. Anal. At. Spectrom., 2018, 33, 2083–2088, DOI: 10.1039/C8JA00263K.

An Improved QZSS Satellite Clock Offsets Prediction Based on the Extreme Learning Machine Method

The Japanese Quasi-Zenith Satellite System (QZSS) has been developed as a GPS (Global Positioning System) complementary system to improve positioning accuracy in the Asia-Pacific region. However, the accuracies of ultra-rapid predicted clocks are not high enough for real-time applications, so it is still a challenge problem. This article focuses on the clock predictions of the new QZSS constellation. Based on the QZSS clock periodic characteristics, an improved clock prediction method combining the spectrum analysis model (SAM) and extreme learning machine (ELM) is proposed with abbreviation as iELM. The key parameters of iELM are selected carefully including the number of hidden layer nodes and activation function. Further, the input length of the iELM network is optimized and discussed thoroughly. For the purpose of assessing the performance of the proposed algorithm, the clock prediction accuracies are compared among GBU-P (the ultra-rapid predicted orbits/clocks provided by GFZ (Deutsches GeoForschungsZentrum)), SAM and iELM methods. It is demonstrated that iELM keeps in a high level of accuracy below 1.0ns with the predicted time increasing from 0 to 18h, and a little larger during the next six hours. And the QZO (Quasi-Zenith Orbit) satellites perform better than GEO (Geostationary Earth Orbit) satellite. Furthermore, precise point positioning (PPP) for both static and kinematic modes are experimentally studied for 13 IGS (International GNSS (Global Navigation Satellite System) Service) MGEX (Multi-GNSS Experiment) stations in the longitude range between 100°E and 180°E. In the static PPP mode, the iELM method is verified to be effective for the GPS/QZSS constellation as positioning accuracy is improved by 28.3%, 57.7% and 47.4% on average in the east, north and up component with respect to GPS/QZSS GBU-P results. Nearly 70.0% stations achieve sub-decimeter accuracy in the vertical component. As for the kinematic PPP, the iELM method based on GPS/QZSS observations performs much better than the others for shorter convergence time and better positioning accuracy. Compared with GPS/QZSS GBU-P, iELM takes at most a half time to get convergence, and the accuracy is enhanced by 27.6%, 23.7% and 13.9% on average in the east, north and up direction respectively.

Forecasting Monthly Runoff Time Series by Single-Layer Feedforward Artificial Neural Network and Grey Wolf Optimizer

Generally, accurate hydrological forecasting information plays an increasingly important role in promoting the comprehensive benefit of hydropower reservoirs. With satisfying generalization ability and search rate, the extreme learning machine (ELM), a famous single-layer feedforward neural network, has been widely used to address regression and classification problem. However, the standard ELM method often falls into second-best solutions with a high probability due to the random assignments of network parameters. In order to overcome this problem, this article aims at developing a hybrid model for monthly runoff time series forecasting. In the hybrid method, an effective swarm intelligence method, grey wolf optimizer (GWO), is adopted to optimize the input-hidden weights and hidden biases of the ELM method; and then the Moore-Penrose generalized inverse method is adopted to determine the hidden-output weights. The world's largest hydropower reservoir, Three Gorges, is chosen to compare the performances of various forecasting methods. Based on the simulation results, the presented method outperforms several traditional forecasting methods (like artificial neural network and support vector machine) in several quantitative indexes. Thus, a novel alternative is presented to predict the nonlinear hydrological time series in China.

Modeling, Simulation and Optimization of Power Plant Energy Sustainability for IoT Enabled Smart Cities Empowered With Deep Extreme Learning Machine

A smart city is a sustainable and effective metropolitan hub, that offers its residents high excellence of life through appropriate resource management. Energy management is among the most challenging problems in such metropolitan areas due to the difficulty and key role of energy systems. To optimize the benefit from the available megawatt-hours, it is important to predict the maximum electrical power output of a baseload power plant. This paper explores the method of a deep extreme learning machine to create a predictive model that can predict a combined cycle power plant’s hourly full-load electrical output. An intelligent energy management solution can be achieved by properly monitoring and controlling these resources through the internet of things (IoT). The universe of artificial intelligence has produced many strides through deep learning algorithms and these methods were used for data analysis. Nonetheless, for further accuracy, deep extreme learning machine (DELM) is another candidate to be investigated for analyses of the data sequence. By using the DELM approach, a high level of reliability with a minimum error rate is achieved. The approach shows better results compared to previous investigations since previous studies could not meet the findings up to the mark and unable to predict power plant electrical energy output efficiently. During the investigation, it is shown that the proposed approach has the highest accuracy rate of 98.6% with 70% of training (33488 samples), 30% of test and validation (14352 examples). Simulation results validate the prediction effectiveness of the proposed scheme.

Solar power forecasting using robust kernel extreme learning machine and decomposition methods

Solar power forecasting using robust kernel extreme learning machine and decomposition methods

Solar power forecasting using robust kernel extreme learning machine and decomposition methods

This paper proposes empirical mode decomposition (EMD)-based robust kernel extreme learning machine (RKELM) to achieve a precise predicted value of solar power generation in a smart grid environment. The non-stationary historical solar power data is initially decomposed into various intrinsic mode functions (IMFs) using EMD, which are subsequently passed through the proposed robust Morlet wavelet kernel extreme learning machine (RWKELM) for solar power prediction at different time horizons. Further a reduced kernel matrix version of RWKELM is used to decrease the training time significantly without appreciable loss of forecasting accuracy. By implementing the real time data for validation of the proposed method for short term solar power prediction it can be observed that the proposed EMD-based RWKELM outperforms various other methods, in terms of different performance matrices and execution time. The solar power prediction results on experimental data show the lowest error which proves the highest prediction accuracy.

Gas Source Declaration With Tetrahedral Sensing Geometries and Median Value Filtering Extreme Learning Machine

Gas source localization (including gas source declaration) is critical for environmental monitoring, pollution control and chemical safety. In this paper we approach the gas source declaration problem by constructing a tetrahedron, each vertex of which consists of a gas sensor and a three-dimensional (3D) anemometer. With this setup, the space sampled around a gas source can be divided into two categories, i.e. inside (“source in”) and outside (“source out”) the tetrahedron, posing gas source declaration as a classification problem. For the declaration of the “source in” or “source out” cases, we propose to directly take raw gas concentration and wind measurement data as features, and apply a median value filtering based extreme learning machine (M-ELM) method. Our experimental results show the efficacy of the proposed method, yielding accuracies of 93.2% and 100% for gas source declaration in the regular and irregular tetrahedron experiments, respectively. These results are better than that of the ELM-MFC (mass flux criterion) and other variants of ELM algorithms.

OP Performance Prediction for Complex Mobile Multiuser Networks Based on Extreme Learning Machine

Due to the complex and variable environments of mobile communication, the mobile multiuser networks become a hot topic. To process active complex event in mobile multiuser networks, it is important to predict the system performance. In this work, the authors consider the multiuser networks which utilizes transmit antenna selection (TAS). We derive novel closed-form expressions for the outage probability (OP) in terms of the Meijer's G-function. Then, a extreme learning machine (ELM)-based OP performance prediction algorithm is proposed. We use the theoretical results to generate training data. We test back-propagation (BP) neural network, locally weighted linear regression (LWLR), wavelet neural network, ELM, and support vector machine (SVM) methods. Compared with wavelet neural network, SVM, BP neural network, and LWLR methods, the Monte-Carlo results shows that the proposed prediction algorithm can consistently achieve higher OP performance prediction results.

Classification of catenary equipment based on a coupled genetic algorithm-extreme learning machine method

In an online monitoring system for an electrified railway, it is important to classify the catenary equipment successfully. The extreme learning machine (ELM) is an effective image classification algorithm and the genetic algorithm (GA) is a typical optimisation method. In this paper, a coupled genetic algorithm-extreme learning machine (GA-ELM) technique is proposed for the classification of catenary equipment. Firstly, the GA is used to search for optimal features by reducing the initial multi-dimensional features to low-dimensional features. Next, the optimised features are used as the input to the ELM. The ELM algorithm is then used to classify the catenary equipment. In this process, the impacts of the activation function, the number of hidden layer neurons and different models on the performance of the ELM are discussed in turn. Finally, the proposed method is compared with traditional methods in terms of classification accuracy and efficiency. Experimental results show that the number of feature dimensions decreases to 58% of the original number and the computational complexity is greatly decreased. Moreover, the reduced features and the few steps of the ELM improve the classification accuracy and speed. Noticeably, when the performance of the GA-ELM method is compared with that of the ELM method, the classification accuracy rate is 93.33% compared with 85.83% and the time consumption is 2.25 s compared with 8.85 s, respectively. That is to say, the proposed method not only decreases the number of features but also increases the classification accuracy and efficiency. This meets the needs of a real-time online condition monitoring system.

A Novel Adaptive Multi-View Non-Negative Graph Semi-Supervised ELM

This paper represents a semi-supervised learning framework, which integrates multi-view learning, extreme learning machine (ELM) and graph-based semi-supervised learning. The aim is to expand the scope of adaptation of non-negative sparse graph (NNSG) framework, under a multi-view condition and a non-linear relationship. The proposed multi-view learning method will be adaptive since when data is single-view the framework will degenerate into an embedded framework for NNSG framework. The proposed ELM method also will be adaptive since the number of hidden layer neuron will change with different number of input and output layer neuron. The combination of both proposed methods outperforms traditional graph-based semi-supervised learning, such as flexible manifold embedding (FME) and NNSG framework, which can not establish an affinity matrix for multi-view and can not establish a non-liner model for unknown data. Unlike traditional graph-based semi-supervised learning methods, which only can label propagation and build linear regression models for single or multi-view data, our proposed method has an obvious advantage that is applicable to any single or multi-view data, and builds linear or non-linear models. We provides extensive experiments on four public database in order to evaluate the performance of the proposed method. These experiments demonstrate significant improvement over the state-of-the-art algorithms in label propagation and processing of new data.