Extreme Learning Algorithm Research Articles

Evaluation of the effectiveness of methodological approaches to hierarchical intelligent control of ground-air Ad-Hoc communication network

The article shows the features of intelligent management processes of a new type of ground-air communication networks that are rapidly being implemented. The dynamic nature of the conditions of operation and behavior of communication nodes, both ground and air, causes a rapid increase in the volume of service information necessary to ensure continuous and adaptive management in real time. One of the ways to solve this problem is the redistribution of management tasks at different stages of the management cycle, which is classically divided into the stages of planning, deployment and operational management. On the one hand, the increase in entropy at the planning stage complicates this process, but this approach will increase the probability of making "correct" management decisions from the standpoint of quality management (network metrics). The work investigates a new architecture of a hierarchical intelligent ground-air communication network control system based on the model-free Reinforcement learning algorithm as a Q-learning network agent and FA-OSELM online sequential extreme machine learning algorithms as node-level agents. The IСS model is presented, its adequacy is checked, and the process of its learning at the planning stage on various mobility models is shown. An important feature of the IСS training process is the application of the developed mobility model, disclosed in the article, which describes the interaction processes of communication nodes at a deeper level. The work conducted a study of the representativeness of the training sample, obtained using the developed mobility model, that was carried out relative to the existing ones, and it was determined that despite the smaller volume of the population of the initial data, it was possible to ensure a better quality of resource management.

OUTPUT POWER ESTIMATION OF A PHOTOVOLTAIC PANEL BY EXTREME LEARNING MACHINE

In this study, the output power of a PV panel under different operating conditions was estimated with the help of an extreme learning algorithm (ELM). For this purpose, a PV panel with a power of 180W was installed, and the open circuit voltage, short circuit current, panel temperature, and solar radiation of this panel were measured and recorded at regular intervals. A total of 75 measurement data were obtained. The maximum power of the panel was calculated using the open circuit voltage and short circuit current information. While panel temperature and solar radiation were given as inputs to the regression model of the PV panel based on ELM, the output of the regression model was taken as the maximum power of the PV panel. To improve the prediction accuracy of ELM, the number of input neurons of ELM and the type of activation function used in the hidden layer were determined by trial and error method. The generated PV data set is separated into training and testing sets. The performance of the method was examined with the 5-fold cross-validation method. For this purpose, the dataset was divided into 5 equal parts. One of these parts was used for testing the ELM and the remaining four sets were used for training the ELM, and this was done by changing the test set each time. Thus, the network was trained and tested 5 times with different sets, and the test result of the network was obtained by averaging the sum of the performances of all test functions. Regression results obtained from ELM are given for different numbers of hidden layer neurons and different types of activation functions in the hidden layer. The best prediction result of ELM was obtained for the case where the hidden layer activation function was tangent sigmoid and the number of hidden layer neurons was 20. From the results obtained, it has been seen that ELM predicts the output power of the PV panel with very high accuracy. It is concluded that ELM is a useful tool for estimating the PV panel output power.

A sustainable health and educational goal development (SHEGD) prediction using metaheuristic extreme learning algorithms

The United Nations established the 17 Sustainable Development Goals (SDGs) in 2015 to address issues like gender equality, clean water, health, education, and hunger by 2030. Of the 17 SDGs, health and education have an outsized impact on countries’ socioeconomic development, so providing insights into progress on these two goals is crucial. Machine learning can help solve many real-world problems, including working towards the SDGs. This paper proposes using a metaheuristic ensemble of Cat Swarm Optimization algorithms with Feed Forward Extreme Learning Machines, called Sustainable Health And Educational Goal Development (SHEGD) Prediction, to effectively contribute to countries’ economic growth by achieving health and education SDGs through machine learning. The model is assessed using UN SDG datasets and performance metrics like accuracy, precision, recall, specificity, and F1-score. Comparisons to other machine learning models demonstrate this model's superiority in designing a recommendation system for progressing towards the health and education SDGs. The proposed model outperforms the other approaches, proving its value for an SDG recommendation system design.

Breast Cancer Diagnosis Using Virtualization and Extreme Learning Algorithm Based on Deep Feed Forward Networks

One of the leading causes of death for women worldwide is breast cancer. Early detection and prompt treatment can reduce the risk of breast cancer-related death. Cloud computing and machine learning are crucial for disease diagnosis today, but they are especially important for those who live in distant places with poor access to healthcare. While machine learning-based diagnosis tools act as primary readers and aid radiologists in correctly diagnosing diseases, cloud-based technology can also assist remote diagnostics and telemedicine services. The promise of techniques based on Artificial Neural Networks (ANN) for sickness diagnosis has attracted the attention of several re-searchers. The 4 methods for the proposed research include preprocessing, feature extraction, and classification. A Smart Window Vestige Deletion (SWVD) technique is initially suggested for preprocessing. It consists of Savitzky-Golay (S-G) smoothing, updated 2-stage filtering, and adaptive time window division. This technique separates each channel into multiple time periods by adaptively pre-analyzing its specificity. On each window, an altered 2-stage filtering process is then used to retrieve some tumor information. After applying S-G smoothing and integrating the broken time sequences, the process is complete. In order to deliver effective feature extraction, the Deep Residual based Multiclass for architecture (DRMFA) is used. In histological photos, identify characteristics at the cellular and tissue levels in both tiny and large size patches. Finally, a fresh customized strategy that combines a better crow forage-ELM. Deep learning and the Extreme Learning Machine (ELM) are concepts that have been developed (ACF-ELM). When it comes to diagnosing ailments, the cloud-based ELM performs similarly to certain cutting-edge technology. The cloud-based ELM approach beats alternative solutions, according to the DDSM and INbreast dataset results. Significant experimental results show that the accuracy for data inputs is 0.9845, the precision is 0.96, the recall is 0.94, and the F1 score is 0.95.

Novel Statistical Regularized Extreme Learning Algorithm to Address the Multicollinearity in Machine Learning

Novel Statistical Regularized Extreme Learning Algorithm to Address the Multicollinearity in Machine Learning

Application of ELM algorithm-based generative adversarial network model in network security

Abstract To improve the efficiency and accuracy of network intrusion discrimination, this paper introduces intrusion detection techniques in a generative adversarial network model. Firstly, a basic framework of a generative adversarial network is constructed. Secondly, the generative adversarial network is trained, and the training process is analyzed to find the data discrimination point in the network. Finally, ELM (Extreme Learning) algorithm is introduced at this discriminating point. The output weight matrix is derived using the minimization square loss function and least squares regression to improve the intrusion discrimination accuracy and intrusion cracking rate in the generative adversarial network, improving network security. To verify the security of the ELM algorithm, this paper simulates the intrusion of the constructed network model, and the results show that the intrusion detection accuracy of the generative adversarial network model based on the ELM algorithm can reach 100%, which is higher than that of DCGAN network 19% and LSGAN network 23%, respectively. The intrusion cracking rate of its layer 5 neural network can reach 92% at the second 2.5 seconds of the simulated intrusion. From the above results, it is clear that the generative adversarial network model based on the ELM algorithm can accurately detect and efficiently crack the intrusion to improve the network security performance.

Development of ecological management system for planted forest based on ELM deep learning algorithm

Plantations play a central and lever role in maintaining the ecological balance of the earth, maintaining the overall function of the terrestrial ecosystem, and promoting the coordinated development of economic society and ecological construction. In order to strengthen the ecological management of plantation forests and improve the ecological level of forest region, the C/S framework is taken as the basic structure, and the programming mode of business model-user interface controller is used, on J2EE platform. The ecological management system of a planted forest is constructed by the evaluation module, the principal component comprehensive analysis module of ecological function value and the demand prediction module of planted forest based on extreme learning machine and deep learning algorithm, and runs under the support of windows system, oracle 15G and above database software. The indexes and factors affecting the ecological function of plantation forests were evaluated and analyzed, and the final management decision was given by the prediction module. The results showed that the plant density significantly affected plant biomass, organic carbon storage, water content and nutrient accumulation, and the comprehensive evaluation indexes of four ecological functions increased from 32.69, 31.84, 33.71 and 35.46 to 86.18, 89.46, 89.83 and 88.76, respectively. Although the degree of influence of the system on lemon strip plants, herbaceous plants, surface litter and soil varies, it still has good feasibility, effectiveness and practicality, and can assist the scientific ecological management of artificial plantation forests.

Using a Hybrid Neural Network and a Regularized Extreme Learning Machine for Human Activity Recognition with Smartphone and Smartwatch.

Mobile health (mHealth) utilizes mobile devices, mobile communication techniques, and the Internet of Things (IoT) to improve not only traditional telemedicine and monitoring and alerting systems, but also fitness and medical information awareness in daily life. In the last decade, human activity recognition (HAR) has been extensively studied because of the strong correlation between people's activities and their physical and mental health. HAR can also be used to care for elderly people in their daily lives. This study proposes an HAR system for classifying 18 types of physical activity using data from sensors embedded in smartphones and smartwatches. The recognition process consists of two parts: feature extraction and HAR. To extract features, a hybrid structure consisting of a convolutional neural network (CNN) and a bidirectional gated recurrent unit GRU (BiGRU) was used. For activity recognition, a single-hidden-layer feedforward neural network (SLFN) with a regularized extreme machine learning (RELM) algorithm was used. The experimental results show an average precision of 98.3%, recall of 98.4%, an F1-score of 98.4%, and accuracy of 98.3%, which results are superior to those of existing schemes.

Covid-19'da Pozitif Vaka Oranı Tahmini İçin Aşırı Öğrenme Makinesi Algoritmaları: Karşılaştırmalı Bir Çalışma

Various pandemics have been recorded in world history until today. The Covid-19 outbreak, which emerged at the end of 2019, has recently been a hot topic in the literature. In this work, extreme learning algorithms are presented as a comparative study for predicting the positive rate for the countries: India, Turkey, Italy, USA and UK. The features to be used in the learning phase are determined with the F-test feature selection method. For each extreme learning approach, results are obtained for each country with the root mean square error evaluation criteria. Accordingly, the radial basis kernel function produces the best estimation results, while the linear kernel function has the highest RMSE. Accordingly, the lowest RMSE value has been obtained for India as 4.17E-03 with the radial basis kernel function based ELM. Also, since Turkey's data contains too many outliers, it has the highest RMSE value (0.015 - 0.035) in linear kernel method among the countries.

Estimation of PM10 and SO2 Substances with Extreme Learning Methods for Analysis of Air Pollution in Amasya Province

Air pollution can lead to various diseases. Especially, Sulfur dioxide (SO2) and Particulate Matter (PM10 ) concentrations in the air directly affect air pollution. Therefore, the estimation of and substances in the air plays an important role in air pollution prediction. In this study, for the data received from the Ministry of Environment and Urbanization air monitoring center, the and pollutants are estimated based on past data with the extreme-learning-adaptive fuzzy extraction system method. Excessive learning algorithms in prediction problems both work fast and can reach low error values. On the other hand, the adaptive neural-fuzzy inference system offers an effective convergence by handling uncertain and incomplete data. Within the scope of the study, estimated and actual values are visualized with figures. The results of extreme learning algorithms are graphed with the error values they reach. Extreme learning-adaptive neural-fuzzy inference system; It is compared with ELM, KELM, CSELM, CDELM, and SELM algorithms. As a result of the comparison, KELM and ELM-ANFIS in terms of RMSE metric for give the best results close to each other with 0.1151 and 0.1155 values, respectively. For , it has been determined that the ELM-ANFIS showed the best performance with 0.0842 RMSE and 0.8171 R2 metric values.

Pressure Image Recognition of Lying Positions Based on Multi-feature value Regularized Extreme Learning Algorithm

Abstract Sleeping postures are one of the indicators for judging sleep quality and preventing sudden diseases. The sleeping postures not only affect people’s sleep quality but also has great significance for the diagnosis of apnea syndrome and bedsores. To realize and recognize the laying positions, this paper researches the regularized extreme learning (RELM) algorithm to analyze the pressure due to lying positions. Based on this algorithm first, the array pressure sensor is used to obtain the back lying posture pressure image, and the image is pre-processed to complete the extraction of multiple feature values (Geometric features, Energy features, and Colour features). Second, the multi-feature values are normalized and finally, these multi-feature values are trained and predicted by the RELM algorithm. In concluding this, the accuracy of lying posture recognition was the highest, achieving 98.75 percent, this is when 1120 datasets of feature values were used as training data and 160 sets as test data while the hidden nodes were 80. RELM algorithm can overcome the problems of extreme learning (ELM) algorithm, such as slow learning speed and local minimum value, and so on. Therefore, this method can be applied in the scenarios of lying posture recognition.

Machine Learning Approach and Model Performance Evaluation for Tele-Marketing Success Classification

Up to the present, various methods such as Data Mining, Machine Learning, and Artificial Intelligence have been used to get the best assess from huge and important data resource. Deep Learning, one of these methods, is extended version of Artificial Neural Networks. Within the scope of this study, a model has been developed to classify the success of tele-marketing with different machine learning algorithms especially with Deep Learning algorithm. Naïve Bayes, C5.0, Extreme Learning Machine and Deep Learning algorithms have been used for modelling. To examine the effect of class label distribution on model success, Synthetic Minority Oversampling Technique have been used. The results have revealed the success of Deep Learning and Decision Trees algorithms. When the data set was not balanced, the Deep Learning algorithm performed better in terms of sensitivity. Among all models, the best performance in terms of accuracy, precision and F-score have been achieved with the C5.0 algorithm.

Extreme Learning Bat Algorithm in Brain Tumor Classification

Brain tumor is considered as an unusual cell that presents and grows in the brain. Similarly, it may lead to cancerous or non-cancerous. So, to improve the survival rate of the patient and to give the best treatment at the earliest, it’s very necessary for early prediction of tumor. Accurate classification of tumor in the brain is important for improving the diagnosis. In accordance with that, various research programs are invited for the better treatment of the patients. Machine Learning (ML) algorithms are applied to help the health associates for the classification of brain tumor and present their diagnosis. This paper focuses primarily on brain tumors of meningioma, Glioma, and pituitary. Moreover, the manual evaluation of Magnetic Resonance Image (MRI) is a difficult process. For accessing MRI brain image in the aspects of its volume, boundaries, detecting tumor size, shape and classification are the challenging tasks. To overcome these difficulties, this paper proposes a novel approach in feature selection using bat algorithm with Extreme Learning Machine (ELM) and for enhancing the accurate classification by Transfer Learning (BA + ELM-TL). Here the data is pre-processed to remove noises; Stationary Wavelet Transforms (SWT) is used to extract the features from the MRI brain image. This paper has collected the dataset from fig share, whole brain atlas and TCGA-GBM data set. Therefore, it is proved that 92.6% is the accuracy of Bat algorithm, 90.4% for Extreme Learning algorithm and 98.87% for BA + ELM-TL.

Fast and robust online-learning facial expression recognition and innate novelty detection capability of extreme learning algorithms

Fast and robust online-learning facial expression recognition and innate novelty detection capability of extreme learning algorithms

Wearable Smart Band for American Sign Language Recognition With Polymer Carbon Nanocomposite-Based Pressure Sensors

The conventional camera-based systems and electronic gloves for gesture recognition are limited by the influence of lighting conditions, occlusions, and movement restrictions. A wearable smart band with integrated nanocomposite pressure sensors has been developed to overcome these shortcomings. The sensors consist of homogeneously dispersed carbon nanotubes in a polydimethylsiloxane polymer matrix prepared by an optimized synthesis process. The sensor band can actively monitor contractions/relaxations of muscles in the arm due to the sensor's high sensitivity in the low forces and stability. The band has eight sensors placed on a stretchable adhesive textile material and connected to a data logger with a multiplexed sensor interface and wireless communication capabilities. The novel smart band was validated by measurements on ten subjects to perform numerical gestures in American sign language from 0 to 9 with ten trials each. The data were recorded at 100 Hz, and a total of 100 datasets were generated for each subject. By feeding the datasets to an extreme machine learning algorithm that selects features, weights, and biases to classify the gestures, an overall gesture recognition accuracy of 93% could be achieved.

Investigations on detection and prevention of primary user emulation attack in cognitive radio networks using extreme machine learning algorithm

Cognitive radio is a competent technology which resolves the spectrum scarcity crisis in a network by allocating free spectrum dynamically and establishes coexistence between the users of the network, without interfering with the incumbent transmission. Primary user emulation attack (PUEA) is one of the major threats to the spectrum sensing performance that decreases the spectrum access probability of the unlicensed users of the network. In this paper Detection and prevention of PUEA is realized proficiently using Time-Distance with signal Strength Evaluation (TDSE) and Extreme Machine Learning (EML) algorithm. TDSE implementation makes it feasible to identify the malicious PUE attacker under mobile user stipulation and EML algorithm provides a swift comparative decision regarding the malicious attack. The technique is also proven to improve the sensing ability, energy efficiency and brings down the overall delay of the system, illustrating a considerable improvement in the overall network performance and spectrum utilization in a cognitive radio network.

Detection and diagnosis of brain tumors‐framework using extreme machine learning and CANFIS classification algorithms

AbstractIn this paper, brain tumors are detected and diagnosed using machine learning approaches in brain magnetic resonance imaging (MRI), which has many real time clinical applications. Noise variations in brain images are detected and removed using index filter, which is proposed in this paper. Brain images devoid of noise content are in spatial domain format, which are not suitable for further feature extraction process. Hence, there is a need for converting all the spatial pixels into multi orientation pixels. In this paper, Gabor transform is used for spatial into multi oriented image conversion. The noise filtered images are transformed into multi orientation‐based brain image using Gabor transform method. Then, the hybrid features which are the integration of statistical and texture features (GLCM, gray level co‐occurrence matrix, and LDP, local derivative pattern), are computed from this transformed brain image. These computed features are classified using extreme machine learning (EML) approach, which categorizes the source brain image as normal or abnormal. Then, the segmented tumor regions are diagnosed using co‐active adaptive neuro fuzzy inference system (CANFIS) classifier, which classifies the segmented regions as mild or severe. The proposed tumor detection and diagnosis methods are applied and tested on the brain images which are available as open access dataset. The performance of the proposed brain tumor detection method is analyzed in terms of sensitivity, specificity, and accuracy with classification rate.

Accuracy assessment of extreme learning machine in predicting soil compression coefficient

The compression coefficient (Cc) is an important soil mechanical parameter that represents soil compressibility in the process of consolidation. In this study, a machine learning derived model, namely extreme learning algorithm (ELM), was used to predict the Cc of soil. A total of 189 experimental results were used and randomly divided to construct the training and testing parts for the development and validation of ELM. Monte Carlo approach was applied to take into account the random sampling of samples constituting the training dataset. A number of 13 input parameters reflecting the experiment were used as the input variables to predict the output Cc. Several statistical criteria, such as mean absolute error (MAE), root mean square error (RMSE), correlation coefficient (R) and the Monte Carlo convergence estimator were used to assess the performance of ELM in predicting the Cc of soil. The results showed that ELM had a strong capacity to predict the Cc of soil, with the R value > 0.95. The convergence of results, as well as the capability of ELM were fully investigated to understand the advantage of using ELM as a predictor.

Monthly runoff forecasting via an improved extreme learning machine

Generally, monthly runoff prediction is of great importance for effective water resource planning and management. Extreme learning machine (ELM) is a novel training tool for the famous single layer feed-forward neural network. Due to the satisfying performance, ELM is chosen for monthly runoff forecasting in this research. Nevertheless, it is unfortunately found that ELM easily falls into local optima in practice because the randomly-determined computational parameters remain unchanged during the learning process. Thus, this paper tries to develop an improved extreme learning algorithm (IELM) where the evolutionary algorithm is used to search for satisfying computational parameters while the Moore-Penrose generalized inverse method is used to determine the output weights. The IELM method is applied to forecast the monthly runoff of Hongjiadu reservoir in southwest China. The results show that the proposed method outperforms several traditional algorithms with respect to the performance indicators. Thus, this paper provides a new and effective artificial intelligence approach for the monthly runoff forecasting.

Research on network intrusion detection security based on improved extreme learning algorithms and neural network algorithms

In order to improve the ability of network fuzzy intrusion detection, a network intrusion detection method based on improved extreme learning algorithm and neural network algorithm is proposed to improve the security of the network. ARMA and other linear detection methods are used to construct the network intrusion signal model, and the nonlinear time series and chaos analysis methods are used to extract the feature of network intrusion and big data information analysis. The limit learning method is used for active detection of network intrusion; the adaptive learning method is used for iterative analysis of network intrusion detection, and the correlation characteristic decomposition method is used to improve the convergence of network intrusion detection. The fuzzy neural network algorithm is used to classify the network intrusion features to improve the intrusion detection performance. The simulation results show that this method has high accuracy and strong anti-jamming ability; it has good application value in network security.