Search

The implementation of novel machine learning models can contribute remarkably to simulating the degradation of concrete due to environmental factors. This study considers the sulfuric acid corrosive factor in wastewater systems to simulate concrete mass loss using five machine learning models. The models include three different types of extreme learning machines, including the standard, online sequential, and kernel extreme learning machines, in addition to the artificial neural network, classification and regression tree model, and statistical multiple linear regression model. The reported values of concrete mass loss for six different types of concrete are the target values of the machine learning models. The input variability was assessed based on two scenarios prior to the application of the predictive models. For the first assessment, the machine learning models were developed using all the available cement and concrete mixture input variables; the second assessment was conducted based on the gamma test approach, which is a sensitivity analysis technique. Subsequently, the sensitivity analysis of the most effective parameters for concrete corrosion was tested using three different approaches. The adopted methodology attained optimistic and reliable modeling results. The online sequential extreme learning machine model demonstrated superior performance over the other investigated models in predicting the concrete mass loss of different types of concrete.

With the development of “Internet+”, online auction platforms of used cars have emerged a lot. As a typical representative of the continuous purchase environment, online sequential auction of used cars faces many uncertainties, including uncertain revenue and risk. To describe them, adopting fuzzy theory to create mean-variance model to estimate the revenue and risk is showed in this paper. Moreover, three types of sellers, aggressive, conservative and rational sellers are analyzed respectively, and strategy models are built, where the multi-criteria optimal function for the latter one is adapted Cobb-Douglas production function. Then, a genetic algorithm based on fuzzy simulation is proposed through integrating the fuzzy simulation and 0-1 genetic algorithm, which can solve the models validly. Lastly, the practical example from Guazi website shows the optimal strategies derived by models can meet sellers’ demands, especially goals of both higher revenue and lower risk for rational sellers, which proves practicability of the model and validity of algorithm.

We investigate monopoly profit under a new online tying strategy, namely sequential bundling. This method allows customers to buy additional products at discounted prices immediately after purchasing one or some of the available products. This strategy has been practiced by Walmart and others but has not been modelled to date. We use microeconomics analysis to conduct a comparison of the gained profit with the three commonly used bundling strategies: no bundling, pure bundling and mixed bundling. The main result shows that the sequential bundling strategy yields higher profits in comparison to all three classic strategies. In particular, for the symmetric cost case, our model provides a useful tool for practitioners such as common online shops.

The characteristics of the data stream have brought enormous challenges to classification algorithms. Concept drift is the most concerning characteristics, and developed classification algorithms must tackle the concept drift problem. Therefore, Extreme Learning Machines (ELM) based algorithms have been developed to respond to the characteristics of the data stream. However, due to randomly assigned input layer weights, ELM based algorithms have encountered problems such as producing inconsistent outputs, generating ill-conditioned matrix, and mapping the inputs to the worst representative space. To overcome these problems, this paper aims to build a stable and well-constructed classifier that responds to the requirements of the data stream by considering all characteristics. A novel data stream classification approach based online sequential ELM (OS-ELM) with unsupervised feature representation learning (UFROS-ELM) and ensemble UFROS-ELM approach based on majority learning is presented in this paper. UFROS-ELM is a modification of the OS-ELM with ELM-AE and concept drift mechanism. ELM-AE is utilized for computing the best discriminative input weights of the classifier. The classifier is then initialized by using the determined weights, first chunk, and OS-ELM algorithm. When a new data chunk arrives, the classifier firstly searches any concept drift occurrence. If it is detected, ELM-AE is utilized to reconstruct the classifier to adapt to the changes. Otherwise, the classifier is sequentially updated updates by processing the current chunk. The results are achieved on the well-known real and artificial data sets and compared with state-of-the-art data stream classification algorithms. The experimental studies demonstrate the achievements of the proposed approaches.

Indoor localization technology has recently attracted an increasing attention in research, among which radio frequency identification (RFID) technology has become a preferred solution due to its advantages in non-line of sight, non-contact and rapid identification. However, in the practical RFID indoor localization application scenarios, the RFID readers need to keep working for an extensive time. When some readers malfunction and fail to be repaired in time, the existing algorithms usually cannot maintain the accuracy of the original localization system. In this paper, we propose an RFID reader-fault-adaptive localization algorithm based on online sequential fuzzy broad learning system, called RF-OSFBLS algorithm. The RF-OSFBLS algorithm improves the fuzzy broad learning system (FBLS) with the ability of online sequential learning (OSFBLS) by using the updating algorithm, so that it can process data streams that continue to arrive in the environment. Meanwhile, RF-OSFBLS algorithm proposes RFID reader-fault-adaptive strategy by introducing a transformation matrix, which can process subsequent data streams when reader fault occurs. We have carried out experiments to study the influence factors and validate the performance, both the simulation and realistic experiment results show that our proposed RF-OSFBLS algorithm can achieve better positioning effect and maintain a relatively high accuracy in the dynamically changing and reader-fault environment.

Because it is difficulty to classify level of fabric wrinkle, this paper proposes a fabric winkle level classification model via online sequential extreme learning machine based on improved sine cosine algorithm (SCA). The SCA has excellent global optimization ability, can explore different search spaces, and effectively avoid falling into local optimum. Because the initial population of SCA will have an impact on its optimization speed and quality, the SCA is initialized by differential evolution (DE) to avoid local optimization, and then the output weight and hidden layer bias are optimized; that is, the improved SCA is used to select the optimal parameters of the online sequential extreme learning machine (OSELM) to improve the generalization performance of the algorithm. To verify the performance of the proposed model DE-SCA-OSELM, it will be compared with other algorithms using a fabric wrinkles dataset collected under standard conditions. The experimental results indicate that the proposed model can effectively find the optimal parameter value of OSELM. The average classification accuracy increased by 6.95%, 3.62%, 6.67%, and 3.34%, respectively, compared with the partial algorithms OSELM, SCAELM, RVFL and PSOSVM, which meets expectations.

Online sequential monitoring of the incidence rates of chronic or infectious diseases is critically important for public health. Governments have invested a great amount of money in building global...

This chapter presents a methodology for the generation of paths from surveillance ground maps. The methodology detects and monitors texture patterns across consecutive images using an online sequential algorithm that creates piecewise linear approximation of the monitored trajectory. It takes under consideration the local structure of the patterns in order to extract a local path. The overall global path is sequentially formed from each segment. We present the preprocessing steps used to extract patterns of interest from an image and the results from two case study examples. The methodology has applications in the autonomous path planning of aerial vehicles following a specified structure, such as that of a river or road.

Online Sequential Extreme Learning Machine (OS-ELM) has been confirmed by numerous studies to be an effective algorithm for online learning scenarios. However, we found that some parameters of OS-ELM are randomly assigned and remain unchanged in the subsequent learning process, which leads to great instability in the model performance in practice. To alleviate this problem, we propose a novel ensemble OS-ELM algorithm (EOS-ELM-R) for solving air quality prediction problems. EOS-ELM-R uses multiple distribution functions to initialize the random parameters of the base OS-ELM models and its final output is the average of the predictions of these base models. Extensive experimental results on two real-world air quality prediction problems show that EOS-ELM-R is effective, and it can achieve better generalization capabilities than similar algorithms.

This paper proposes a density-based semi-supervised online sequential extreme learning machine (D-SOS-ELM). The proposed method can realize online learning of unlabeled samples chunk by chunk. Local density and distance are used to measure the similarity of patterns, and the patterns with high confidence are selected by the ‘follow’ strategy for online learning, which can improve the accuracy of learning. Through continuous patterns selection, the proposed method ultimately achieves effective learning of unlabeled patterns. Furthermore, using local density and relative distance can effectively respond to the relationship between patterns. Compared with the traditional distance-based similarity measure, the ability to deal with complex data is improved. Empirical study on several standard benchmark data sets demonstrates that the proposed D-SOS-ELM model outperforms state-of-art methods in terms of accuracy.

Online sequential extreme learning machine (OS-ELM) and its variants provide a promising way to solve data stream problems, but most of them do not take the timeliness of the problems into account, which may degrade the performance of the model. The main reason is that these algorithms are unable to adapt to the latest data accordingly when the distribution of the data stream changes. To mitigate this limitation, the forgetting factor is introduced into the relevant models, which is used to balance the relative importance of past data and new data when necessary. However, there is no efficient way to set the forgetting factor properly so far. In this paper, we have developed a novel updating strategy for setting the forgetting factor and proposed a dynamic forgetting factor based OS-ELM algorithm (DOS-ELM). In the sequential learning phase of DOS-ELM, the forgetting factor can be adjusted dynamically according to the change degree of the model accuracy in each learning epoch. This updating process does not require setting any parameters artificially and thus greatly improves the flexibility of the model. The experimental results on ten classification problems, five regression problems, one time-series problem show that DOS-ELM can deal well with both stationary and non-stationary data stream problems. In addition, we have extended DOS-ELM to an online deep model named ML-DOS-ELM, which can handle more complex tasks such as the face recognition problem and the handwritten digit recognition problem. Our experimental evaluations show that both DOS-ELM and ML-DOS-ELM can achieve higher prediction accuracy compared to the other similar algorithms.

In this paper, an online sequential learning algorithm known as online sequential extreme learning machine (OS ELM) is applied to simulate the power output of a wind turbine. The OS ELM is used both in 1-by-1 and chunk-bychunk mode and the results are compared with batch learning algorithms, namely Back Propagation (BP) and Extreme Learning Machine (ELM) algorithm. Different activation functions such as Sigmoidal, Sin, Radial Basis Function (RBF) and Hardlim have been used in OS ELM to decide upon most optimal function. It has been found that OS ELM with fixed chunk size of 50-by-50 and sigmoidal activation function with training time of 0.080s, Root Mean Square Error (RMSE) of 1.96%, prediction accuracies on training and test data of 100% and 99.95 % respectively, is best suited for wind power modelling and simulation applications, where the data arrives in a sequential manner.

The uncertainty of demands brings challenges for the private cloud providers, leading to low utilization of resources during periods of low-demand and low quality of service during periods of peak-demand, which has attracted much attention. In this paper, taking account into both uncertainty of demands and budget constraint, we design an online sequential procurement auctions of residual resources, which helps the busy cloud provider make an irrevocable decision about how to purchase resources during period of uncertain peak-demand. The crucial part of the mechanism is the seller accepting-rule based on a value-density threshold which is learned dynamically from the historical information. Given the condition that all the sellers are myopic, we prove that the mechanism is truthful, budget feasible and individual rational. Furthermore, we obtain the competitive ratio of the proposed mechanism when the demands of the BCP are δ-degree balance. Using real data from parallel computing centers, we construct 60 scenarios in six data settings, in which we compare our mechanism with average budget allocation and offline proportional sharing mechanism, the results show that in more than 85% scenarios the proposed mechanism has better performance than allocation with average budget, and it improves more than 20% valuation on average for the buyer, even if we use the estimate value of balance degree δ.

Hematocrit (HCT) is expressed as the percentage of red blood cells in the whole blood, it is one of the most highly affecting factors which influences the glucose measurement by using handheld device. In this paper, we present an approach for applying the regularized online sequential learning to hematocrit estimation. The input is the transduced current curve which is produced by the chemical reaction during glucose measurement. The experimental results shown that the proposed approach is promising.

The main contribution of this paper is to introduce a new iterative training algorithm for restricted Boltzmann machines. The proposed learning path is inspired from online sequential extreme learning machine one of extreme learning machine variants which deals with time accumulated sequences of data with fixed or varied sizes. Recursive least squares rules are integrated for weights adaptation to avoid learning rate tuning and local minimum issues. The proposed approach is compared to one of the well known training algorithms for Boltzmann machines named “contrastive divergence”, in term of time, accuracy and algorithmic complexity under the same conditions. Results strongly encourage the new given rules during data reconstruction.

Field-programmable gate array (FPGA) has been used as a very effective hardware platform in different research area as it enhances the efficiency of the embedded module. The accessibility of minimized, fast circuitry for a artificial neural networks (ANNs) is the most important and utmost necessity for many critical applications. In this paper, a single layer feed-forward neural network (SLFN) named as online sequential extreme learning machine (OS-ELM) is conferred and realized in digital platform for real-world data classification. The digital employment of OS-ELM supports to form an efficient hardware unit for data classification in real-time as the classifier has high learning speed and promising accuracy. Finally, the digital architecture of OS-ELM is implemented on a Virtex-5 FPGA hardware platform to validate the feasibility, efficacy, and vitality of the proposed classifier in real-time.

The objective of classification in high dimensional biomedical data is to map an input feature space to a predefined class labels with higher classification accuracy and less computational time. Although various machine learning techniques have been developed till now, still it is a challenging issue among researchers to develop accurate classifier for better diagnosis of various diseases. Therefore, in this study a new classifier named as adaptive fruit fly optimization (AFFO) based online sequential extreme learning machine (OSELM) is presented to classify biomedical data like Indian diabetes, Parkinson and Indian liver patient available in UCI repository. The proposed model performance is also compared with various models like AFFO-ELM and OSELM. The performances of these models are validated using some performance measures such as G-mean, sensitivity, F-Score, accuracy, specificity, and precision. The results confirm that AFFO-OSELM performance outperforms over other models used in this study.

Prediction models are essential in dam crack behavior identification. Prototype monitoring data arrive sequentially in dam safety monitoring. Given such characteristic, sequential learning algorithms are preferred over batch learning algorithms as they do not require retraining whenever new data are received. A new methodology using the genetic optimized online sequential extreme learning machine and bootstrap confidence intervals is proposed as a practical tool for identifying concrete dam crack behavior. First, online sequential extreme learning machine is adopted to build an online prediction model of crack behavior. The characteristic vector of crack behavior, which is taken as the online sequential extreme learning machine input, is extracted by the statistical model. A genetic algorithm is introduced to optimize the input weights and biases of online sequential extreme learning machine. Second, the BC a method is proposed to produce confidence intervals based on the improved online sequential extreme learning machine prediction. The improved online sequential extreme learning machine for identifying crack behavior is then built. Third, the crack behavior of an actual concrete dam is taken as an example. The capability of the built model for predicting dam crack opening is evaluated. The comparative results demonstrate that the improved online sequential extreme learning machine can provide highly accurate forecasts and reasonably identify crack behavior.

This study proposes an ensemble differential evolution online sequential extreme learning machine (DE-OSELM) for textile image illumination correction based on the rotation forest framework. The DE-OSELM solves the inaccuracy and long training time problems associated with traditional illumination correction algorithms. First, the Grey–Edge framework is used to extract the low-dimensional and efficient image features as online sequential extreme learning machine (OSELM) input vectors to improve the training and learning speed of the OSELM. Since the input weight and hidden-layer bias of OSELMs are randomly obtained, the OSELM algorithm has poor prediction accuracy and low robustness. To overcome this shortcoming, a differential evolution algorithm that has the advantages of good global search ability and robustness is used to optimize the input weight and hidden-layer bias of the DE-OSELM. To further improve the generalization ability and robustness of the illumination correction model, the rotation forest algorithm is used as the ensemble framework, and the DE-OSELM is used as the base learner to replace the regression tree algorithm in the original rotation forest algorithm. Then, the obtained multiple different DE-OSELM learners are aggregated to establish the prediction model. The experimental results show that compared with the textile color correction algorithm based on the support vector regression and extreme learning machine algorithms, the ensemble illumination correction method achieves high prediction accuracy, strong robustness, and good generalization ability.

In online learning, the contribution of old samples to a model decreases as time passes, and old samples gradually become invalid. Although the Online Sequential Extreme Learning Machine (OS-ELM) can avoid the repetitive training of old samples, invalid samples are still used, which goes against improving the accuracy of an OS-ELM model. The Online Sequence Extreme Learning Machine with Forgetting Mechanism (FOS-ELM) timely discards invalid samples, but it does not consider the differences among valid samples and then has the limitation on boosting the accuracy and generalization. To solve this issue, the Memory Degradation Based OS-ELM (MDOS-ELM) is proposed in this paper. The MDOS-ELM adjusts the weights of the old and new samples in real time by a self-adaptive memory factor, and simultaneously discards invalid samples. The self-adaptive memory factor is determined by two elements. One is the similarity between the new and old samples, and the other is the prediction errors of the current training samples on the previous model. The performance of the proposed MDOS-ELM is validated on both regression and classification datasets which include an artificial dataset and twenty-two real-world dataset. The results demonstrate that the MDOS-ELM model outperforms the OS-ELM and the FOS-ELM models on the accuracy and generalization.

Contextual multi-armed bandit algorithms are an effective approach for online sequential decision-making problems. However, there are limited tools available to support their adoption in the community. To fill this gap, we present an open-source Python library with context-free, parametric and non-parametric contextual multi-armed bandit algorithms. The MABWiser library is designed to be user-friendly and supports custom bandit algorithms for specific applications. Our design provides built-in parallelization to speed up training and testing for scalability with special attention given to ensuring the reproducibility of results. The API makes hybrid strategies possible that combine non-parametric policies with parametric ones, an area that is not explored in the literature. As a practical application, we demonstrate using the library in both batch and online simulations for context-free, parametric and non-parametric contextual policies with the well-known MovieLens data set. Finally, we quantify the performance benefits of built-in parallelization.

As real industrial processes have measurement samples with noises of different statistical characteristics and obtain the sample one by one usually, on-line sequential learning algorithms which can achieve better learning performance for systems with noises of various statistics are necessary. This paper proposes a new online Extreme Learning Machine (ELM, of Huang et al.) algorithm, namely recursive least mean p-power ELM (RLMP-ELM). In RLMP-ELM, a novel error criterion for cost function, namely the least mean p-power (LMP) error criterion, provides a mechanism to update the output weights sequentially. The LMP error criterion aims to minimize the mean p-power of the error that is the generalization of the mean square error criterion used in the ELM. The proposed on-line learning algorithm is able to provide on-line predictions of variables with noises of different statistics and obtains better performance than ELM and online sequential ELM (OS-ELM) while the non-Gaussian noises impact the processes. Simulations are reported to demonstrate the performance and effectiveness of the proposed methods.

Diagnosis of bearings generally plays an important role in fault diagnosis of mechanical system, and machine learning has been a promising tool in this field. In many real applications of bearings fault diagnosis, the data tend to be online imbalanced, which means, the number of fault data is much less than the normal data while they are all collected in online sequential way. Suffering from this problem, many traditional diagnosis methods will get low accuracy of fault data which acts as the minority class in the collected bearing data. To address this problem, an online sequential prediction method for imbalanced fault diagnosis problem is proposed based on extreme learning machine. This method introduces the principal curve and granulation division to simulate the flow distribution and overall distribution characteristics of fault data, respectively. Then a confident over-sampling and under-sampling process is proposed to establish the initial offline diagnosis model. In online stage, the obtained granules and principal curves are rebuilt on the bearing data which are arrived in sequence, and after the over-sampling and under-sampling process, the balanced sample set is formed to update the diagnosis model dynamically. A theoretical analysis is provided and proves that, even existing information loss, the proposed method has lower bound of the model reliability. Simulation experiments are conducted on IMS bearing data and CWRU bearing data. The comparative results demonstrate that the proposed method can improve the fault diagnosis accuracy with better effectiveness and robustness than other algorithms.

Autoencoder is a neural network model which extracts the best representing features of data by performing unsupervised prelearning. Existing autoencoder approaches have been developed for applications that can access all data set at a time. However, the existing autocoders fail to satisfy for data stream applications that does not have access to the entire data set at once and and requires instant response. Therefore, this study presents an online sequential autocoder approach for data stream. The proposed autoencoder approach is based on the Extreme Learning Machines Autoencoder (ELM-AE) and adapts ELM-AE to be incremental learning algorithm. Therefore, Online Sequential-ELM (OS-ELM) algorithm which is incremental learning algorithm for data stream is integrated into ELM-AE in this paper. The proposed approach is tested on KddCup99, Poker, Pendigits, Hyperplane and ElecNormNews data stream sets which are commonly used in literature and achieved results demonstrate proposed approach's success.

In this paper, we propose a complete online learning framework for object detection system creatively. The framework efficiently combines Pixel Intensity Comparisons Organized in Decision Trees (PICO) and Local Receptive Fields Based Extreme Learning Machine with Online Sequential Learning Mechanism (OS-ELM-LRF). OS-ELM-LRF is the modified ELM-LRF for which we add the online sequential mechanism. In this framework, PICO is used as the object detector to obtain core candidate regions with high confidence, while OS-ELM-LRF is applied as the object classifier to recognize the specific target. This is an extremely lightweight and efficient online learning framework that can be ported to some embedded devices. To illustrate the effectiveness of this framework, we realize the face recognition system and compare it to the deep-learning-based detection system. Experimental results demonstrate that the proposed object detection framework has not only high recognition accuracy, extremely real-time performance but also remarkable online learning ability, and it can be extended for most object detection tasks in industrial production.

At present, machine learning techniques is frequently used to extraction and/or prediction some information from enormous collected data. Traditional machine learning is performed batch learning like a popular one, Artificial Neural Network (ANN). That means it can’t integrate new information into already trained model but it is retrained from scratch. Online Sequential Extreme Learning Machine (OS-ELM) is a one of online incremental machine learning techniques that can learn and update model from new receive data without to retrain the model. This paper shows the result of comparison experiment between ANN and OS-ELM and found that, the OS-ELM has acceptable performance in situation of few initial data for training and/or statistic properties of data is changing while it working.

Glucose measurement by using handheld devices is applied widely due to their comfortabilities. They are easy to use and can give results quickly. However, the accuracy of measurement results is affected by interferences, in which hematocrit (HCT) is one of the most highly affecting factors. In this paper, an approach for glucose correction based on the neural network is presented. The regularized online sequential learning is utilized for hematocrit estimation. The transduced current curve which is produced by the chemical reaction during glucose measurement is used as an input feature of neural network. The experimental results shown that the proposed approach is promising.

Analytical redundancy technique is of great importance to guarantee the reliability and safety of aircraft engine system. In this paper, a machine learning based aeroengine sensor analytical redundancy technique is developed and verified through hardware-in-the-loop (HIL) simulation. The modified online sequential extreme learning machine, selective updating regularized online sequential extreme learning machine (SROS-ELM), is employed to train the model online and estimate sensor measurements. It selectively updates the output weights of neural networks according to the prediction accuracy and the norm of output weight vector, tackles the problems of singularity and ill-posedness by regularization, and adopts a dual activation function in the hidden nodes combing neural and wavelet theory to enhance prediction capability. The experimental results verify the good generalization performance of SROS-ELM and show that the developed analytical redundancy technique for aeroengine sensor fault diagnosis based on SROS-ELM is effective and feasible.

The online sequential extreme learning machine (OS-ELM) was proposed as an evolution of the Extreme learning machine (ELM) in 2006. The learning algorithm can learn data by fixing or changing the size of the chunk (a block of data). Compared with other online algorithms, it has faster training speed, and the generalization performance is better. The proposed ensemble of neural networks further demonstrates that the use of the neural network sets with multiple consensus schemes has a greater improvement over the stability of a single network. The basic idea of cross-validation is to group the original data sets, part of which is to train as the training set and the other part of which is to verify as the test set, so the algorithm can reduce the over-fitting problem of the data to a certain extent, and extract more valid information in the limited data set. In this paper, a new ensemble of online sequential extreme learning machine based on cross-validation (ENOS-ELM) is proposed. Through application of cross-validation and integration to the training stage, the accuracy of the algorithm in classification is further improved.

The extreme learning machine (ELM) is an analytical algorithm developed to train single-hidden layer feedforward neural network (SLFN). Although the ELM provides a good generalization and a very fast training phase, two shortcomings can be pointed in this approach: it assigns its input weights randomly and it is not able to handle sequential learning. Thereby, the online sequential extreme learning machine (OS-ELM) was proposed in order to tackle the lack of sequential learning on ELMs. Nonetheless, the OS-ELM still assigns the input weights randomly. Thus, in this work, we proposed an approach based on a restricted Boltzmann machine (RBM) to improve the OS-ELM by determining its input weights throughout the training phase. We applied our method to several well-known benchmarks and the results show that the proposed approach is able to improve the OS-ELM for most them.

Hot metal silicon content is an important indicator of measuring the smooth operation of the blast furnace. However, the hot metal silicon content cannot be directly detected online. Hence, this paper establishes a prediction model of the hot metal silicon content based on multi-layer online sequential extreme learning machine (ML-OSELM). ML-OSELM is an online sequential version of multi-layer extreme learning machine (ML-ELM), and it can learn data one-by-one or chunk-by-chunk. It uses online sequential extreme learning machine auto-encoder (OS-ELM-AE) to perform layer-by-layer unsupervised sequential learning. The prediction model of the hot metal silicon content based on ML-OSELM is verified by using the production data of the blast furnace. Simulation results demonstrate that this prediction model can accurately and quickly predict the hot metal silicon content. Moreover, this prediction model can provide efficient decision for subsequent operation of the blast furnace.

Solar energy, which is one of the greatest potential renewable energies, has widely attracted attention in current years. Accurate prediction of solar radiation is the premise of exploiting and utilization of solar energy. However, most of the research works on solar radiation prediction focus on the offline prediction, which is not practical and suitable in real world applications. In order to tackle this issue, in this paper, we implement a new kind of machine learning algorithm, online sequential extreme learning machine (OS-ELM), to realize the real-time prediction of solar radiation. Comparing with existing batch learning algorithms for solar radiation prediction, OS-ELM can handle sequentially coming data one-by-one or chunk-by-chunk with fixed or varying chunk size. Its online learning capability makes it possible to reflect and adapt to the environmental changes in a timely manner.

Online sequential extreme learning machine (OS-ELM) can learn the data one-by-one or chunk-by-chunk with the fixed or varying chunk size. It was proposed by Liang et al. is a faster and more accurate algorithm as compared to other online learning algorithms. However, besides the advantages of OS-ELM machine, the original OS-ELM algorithm also introced some issues; first, the improved OS-ELM algorithms need to be network structure adjustment to improve learning promance; second, OS-ELM algorithm learning with stability will affect its generalization ability. For such reasons, in this paper we propose a survey of OS-ELM algorithm with the development of history and the latest results of researching which can hopefully support researchers in the furture.

Traditional pattern recognition algorithms have limitations including slow training speed and low recognition accuracy in practical engineering applications. In this paper, a new method based on Online Sequential Extreme Learning Machine (OS-ELM) is proposed. Data samples have been obtained from PD experiment of real transformer based on Ultra High Frequency (UHF) detection method. In addition, OS-ELM is compared with Extreme Learning Machine (ELM), Support Vector Machine (SVM) and Back Propagation Neural Network (BPNN) in both recognition accuracy and performance aspects. The results show that OS-ELM is not only much faster in learning speed, but also more excellent in recognition accuracy, thus more suitable for engineering applications with large volume of data samples.

Many real world applications are of time-varying nature and an online learning algorithm is preferred in tracking the real-time changes of the time-varying system. Online sequential extreme learning machine (OSELM) is an excellent online learning algorithm, and some improved OSELM algorithms incorporating forgetting mechanism have been developed to model and predict the time-varying system. But the existing algorithms suffer from a potential risk of instability due to the intrinsic ill-posed problem; besides, the adaptive tracking ability of these algorithms for complex time-varying system is still very weak. In order to overcome the above two problems, this paper proposes a novel OSELM algorithm with generalized regularization and adaptive forgetting factor (AFGR-OSELM). In the AFGR-OSELM, a new generalized regularization approach is employed to replace the traditional exponential forgetting regularization to make the algorithm have a constant regularization effect; consequently the potential ill-posed problem of the algorithm can be completely avoided and a persistent stability can be guaranteed. Moreover, the AFGR-OSELM adopts an adaptive scheme to adjust the forgetting factor dynamically and automatically in the online learning process so as to better track the dynamic changes of the time-varying system and reduce the adverse effects of the outdated data in time; thus it tends to provide desirable prediction results in time-varying environment. Detailed performance comparisons of AFGR-OSELM with other representative algorithms are carried out using artificial and real world data sets. The experimental results show that the proposed AFGR-OSELM has higher prediction accuracy with better stability than its counterparts for predicting time-varying system.

Cast shadow detection and removal is one of the key problems in vision-based systems for accurate and robust segmentation of moving objects. This paper proposes a co-training-based adaptive method for detecting moving shadows in video sequences. Shadow detection based on static methods cannot adapt to changing environment such as gradual illumination changes. In order to solve this problem, we have proposed an online sequential extreme learning machine (OS-ELM)-based semi-supervised technique for moving cast shadow detection. Online learning of OS-ELM is much faster and provides better generalization performance compared to other popular online learning algorithms. First, we extracted color, texture, gradient, and image patch similarity features using a background model and input video frame, which are useful for discriminating moving shadows and objects. Co-training scheme is used for online updating of the OS-ELM classifier in order to adapt to the dynamic environment. Experimental results on different benchmark video sequences shows that the proposed method performs better shadow detection and discrimination compared with other methods.

The online sequential extreme learning machine (OS-ELM) has been used for training without retraining the ELM when a chunk of data is received. However, OS-ELM may be affected by an improper number of hidden nodes settings which reduces the generalization of OS-ELM. This paper addresses this problem in OS-ELM. A new structural tolerance OS-ELM (STOS-ELM), based on the Householder block exact inverse QRD recursive least squares algorithm having numerical robustness is proposed. Experimental results conducted on four regressions and five classification problems showed that STOS-ELM can handle the situation when the network is constructed with an improper number of hidden nodes. Accordingly, the proposed STOS-ELM can be easily applied; the size of the hidden layer of ELM can be roughly approximated. If a chunk of data is received, it can be updated in the existing network without having to worry about the proper number of given hidden nodes. Furthermore, the accuracy of the network trained by STOS-ELM is comparable to that of the batch ELM when the networks have the same configurations. STOS-ELM can also be applied in ensemble version (ESTOS-ELM). We found that the stability of STOS-ELM can be further improved using the ensemble technique. The results show that ESTOS-ELM is also more stable and accurate than both of the original OS-ELM and EOS-ELM, especially in the classification problems.

A novel particle swarm optimization based selective ensemble (PSOSEN) of online sequential extreme learning machine (OS-ELM) is proposed. It is based on the original OS-ELM with an adaptive selective ensemble framework. Two novel insights are proposed in this paper. First, a novel selective ensemble algorithm referred to as particle swarm optimization selective ensemble is proposed, noting that PSOSEN is a general selective ensemble method which is applicable to any learning algorithms, including batch learning and online learning. Second, an adaptive selective ensemble framework for online learning is designed to balance the accuracy and speed of the algorithm. Experiments for both regression and classification problems with UCI data sets are carried out. Comparisons between OS-ELM, simple ensemble OS-ELM (EOS-ELM), genetic algorithm based selective ensemble (GASEN) of OS-ELM, and the proposed particle swarm optimization based selective ensemble of OS-ELM empirically show that the proposed algorithm achieves good generalization performance and fast learning speed.

Accurate and fast prediction of nonstationary time series is challenging and of great interest in both practical and academic areas. In this paper, an online sequential extreme learning machine with new weighted strategy is proposed for nonstationary time series prediction. First, a new leave-one-out (LOO) cross-validation error estimation for online sequential data is proposed based on inversion of block matrix. Second, a new weighted strategy based on the proposed LOO error estimation is proposed. This strategy ranks the samples’ importance by means of the LOO error of each new added sample and then assigns various weights. Performance comparisons of the proposed method with other existing algorithms are presented based on chaotic and real-world nonstationary time series data. The results show that the proposed method outperforms the classical ELM and OS-ELM in terms of generalization performance and numerical stability.

This paper proposes a novel recurrent multi-steps-prediction model call Recurrent Kernel Online Sequential Extreme Learning Machine with Surprise Criterion (SC-RKOS-ELM). This model combines the strengths of Kernel Online Sequential Extreme Learning Machine (KOS-ELM), the characteristics of surprise criterion and advantages of recurrent multi-steps-prediction algorithm to unleash the restriction of prediction horizon and reduce the computation complexation of the learning part. In the experiment, we employ two synthetic and two real-world data sets, including Mackey-Glass, Lorenz, palm oil price and water level in Thailand, to evaluate Recurrent Online Sequential Extreme Learning Machine (ROS-ELM) and Recurrent Kernel Online Sequential Extreme Learning Machine with Fixed-budget Criterion (FB-RKOS-ELM). The results of experiments indicate that SC-RKOS-ELM has the superior predicting ability in all data sets than others.

A novel approach to the fault location estimation and prediction analysis in doubly fed power transmission lines is presented. This approach uses an Artificial Intelligence (AI) technique and technology called the Hierarchical Temporal Memory (HTM). HTM is a state-of-the-art online learning neural/machine learning technique and tool based on a suite of biological plausible algorithms coined the Cortical Learning Algorithm (CLA). The results on using this new technique on an artificially generated fault data for fault location prediction are presented as well. The results show that the proposed approach can indeed lead to superior results when compared to a similar conventional machine learning technique called the Online Sequential Extreme Learning Machine (OS-ELM).

Microscopically small solid particles and liquid droplets suspended in the air, known as particulate matter (PM), may significantly affect not only human health but also urban, natural and agricultural systems. Therefore, it is imperative to keep the concentration levels of these pollutants below harmful thresholds. Forecasting tools based on machine learning have been used to estimate the concentration of PM and other pollutants in the atmosphere. However, PM data are uninterruptedly collected over time, thus producing a stream of data whose distribution may evolve over time. As traditional machine learning techniques do not have mechanisms to handle changes on data distribution at running time, they usually present limited prediction accuracy when facing such scenario. The overall goal of this work is to evaluate whether online sequential learning can improve the estimation accuracy of PM forecasting. To do so, online and offline algorithms based on Extreme Learning Machines (ELM) were compared, in order to evaluate their performance when applied to forecast hourly concentrations of PM. The experiments were performed using real-world data streams of PM concentration from different cities of the State of Sao Paulo, Brazil. The obtained results show not only that online sequential learning approaches lead to smaller mean squared errors but also that the stability of the results is enhanced when such approaches are combined in ensembles.

This paper aims to study the use of a multistep parallel-strategy (MSPS) for long-term estimation of the remaining useful life of the Li-ion batteries. Various extreme learning machines (ELMs) including standard ELM, kernel ELMs and online sequential ELM (OS-ELM) are used along with the parallel strategy for multi-step prognosis. These multistep predictors are trained by means of constant current Li-ion battery datasets. The attained results present the effectiveness of these MSPS techniques in the long-term prediction of the remaining useful life of the Li-ion batteries. Besides, the OS-ELM predictor outperforms other techniques in terms of the root mean square error (RMSE) and the RUL estimation error.

In this paper, a novel method called online sequential extreme learning machine with under-sampling and over-sampling (OSELM-UO) for imbalanced Big data classification is proposed which combines the structures of under-sampling and over-sampling and applies online sequential extreme learning machine as its base model. The novel structure enables OSELM-UO performs well on both minority and majority classes and simultaneously overcomes the issues of information loss and overfitting. Moreover, when the dataset keeps growing, OSELM-UO can be applied without retraining all previous data. Experiments have been conducted for OSELM-UO and several imbalance learning methods over real-world datasets respectively under high imbalance ratio (IR) and large amount of samples and features. Through the analysis of the experimental results, OSELM-UO is shown to give the best results in various aspects.

An integral part of modem day health-care is monitoring the physical activities of human beings. In this paper, we deal with automatic recognition of some daily activities based on signals measured using easily-available smart phones. We present a neural-network based methodology to classify these signals. In contrast to typical conventional techniques we use sequential processing of signals and circumvent pre-processing and feature extraction. In addition, we introduce meta-cognition to reduce the computations required during the training stage. We demonstrate that our approach yields satisfactory recognition accuracies.