Random Vector Functional Link Research Articles

Clothing Image Classification with DenseNet201 Network and Optimized Regularized Random Vector Functional Link

ABSTRACT To ameliorate the precision of clothing image classification, we proposed a clothing image classification method via the DenseNet201 network based on transfer learning and the optimized regularized random vector functional link (RVFL). First, the formula extracts weight’s parameters about DenseNet201 that is pre-trained on the ImageNet dataset for transfer learning, thereby obtaining an incipient network，after that trim this model parameters. The modified network is utilized to pick up the clothing image features output by the DenseNet201’s global average pooling layer. Second, regularization coefficient is introduced to control RVFL’s model complexity and solve the problem of over-fitting. Then, the generated solution vector of aquila optimizer (AO) is produced by marine predators algorithm (MPA). The input weights, biases of hidden layer and renormalization modulus of regularized RVFL are optimized using the improved AO algorithm. Finally, we use the optimized RVFL to assort abstracted fashion graphics traits. We use Accuracy, Macro-F1, Macro-R and Macro-P to assess the algorithm’s ability and compare this algorithm with ResNet50 network, ResNet101 network, DenseNet201 network, InceptionV3 network and different classifiers, which use DenseNet201 as the feature extractor to get the input. From the experimental results, this algorithm proposed has excellent classification power and generalization ability.

Limited information limits accuracy: Whether ensemble empirical mode decomposition improves crude oil spot price prediction?

A reliable crude oil price forecast is important for market pricing. Despite the widespread use of ensemble empirical mode decomposition (EEMD) in financial time series forecasting, the one-time decomposition on the entire time series leads the in-sample data to be affected by the out-of-sample data. Consequently, the forecasting accuracy is overstated. This study incorporates a rolling window into two prevalent EEMD-based modeling paradigms, namely decomposition-ensemble and denoising, to ensure that only in-sample time series is processed by EEMD and used for model training. Given the time-consuming process of stepwise preprocessing and model fitting, two non-iterative machine learning algorithms, random vector functional link (RVFL) neural network and extreme learning machine (ELM), are used as predictors. Hence, we develop the rolling decomposition-ensemble and rolling denoising paradigms, respectively. Contrary to the majority of prior studies, empirical results based on monthly spot price time series for the Brent and West Texas Intermediate (WTI) markets indicate that EEMD plays a weak role in improving crude oil price forecasts when only the in-sample set is preprocessed. This is compatible with the weak form of the efficient market hypothesis (EMH). Nevertheless, the suggested rolling EEMD-denoising model has an advantage over other employed models for long-term forecasting.

A novel global solar exposure forecasting model based on air temperature: Designing a new multi-processing ensemble deep learning paradigm

The total quantity of solar energy falling on a horizontal plane surface is the global solar exposure (GSE, i.e., total solar energy). Precise forecasting of GSE is important in many fields such as renewable energy, agriculture, and public health, particularly by the limited hydro-meteorological time series information. This research aims to develop an advanced multi-processing deep learning (DL) paradigm to forecast weekly GSE based on maximum (Tmax) and minimum (Tmin) air temperatures as the drivers at Brisbane and Perth airport stations in eastern and western Australia during 2000 to 2022. The proposed model was comprised of an extra tree feature selection (FS) integrated with two novel decomposition techniques, namely time-varying filtering-based empirical mode decomposition (TVF-EMD), and empirical wavelet transform (EWT), and a powerful ensemble deep random vector functional link (ED-RVFL) approach. To validate the main model, the RVFL, bidirectional long-short term memory (Bi-LSTM), and bagged regression tree (Bagging) machine learning (ML) models were examined in hybrid and standalone counterpart frameworks. First, the extra-tree FS determined the significant lags of the predictors based on an importance benchmark criterion. Then, by applying the optimal gained lags to the feeding models, all of the original predictors were decomposed using TVF-EMD and EWT univariate feature extraction. The final forecast was computed by aggregating all the individual forecasts of the intrinsic mode functions (IMFs) and residual components. In addition, eight statistical indicators (including coefficient of correlation: R, root mean square error: RMSE, Kling-Gupta efficiency: KGE, index of agreement: IA, uncertainty coefficient with 95% confidence level: U95%, mean absolute percent error: MAPE, Nash-Sutcliffe efficiency: NSE, and mean absolute error: MAE) and several graphical methods were utilized to evaluate the performances of the models. The modeling results indicated that the ED-RVFL-TVF-EMD (R = 0.9665, RMSE = 1.9193 MJ/m2, and KGE = 0.9565 for the Perth airport station) and ED-RVFL-EWT (R = 0.9218, RMSE = 1.9708 MJ/m2, and KGE = 0.8552 for the Brisbane station) outperformed all other models, followed by RVFL, Bi-LSTM, and Bagging in a hybrid format. With the high predictive robustness, both decomposition-based frameworks can be useful for solving energy forecasting problems. The new modeling approach developed in this study can provide more precise forecasts for decision-makers to better address climate change, agriculture, and energy crises.

Application of Advanced Optimized Soft Computing Models for Atmospheric Variable Forecasting

Precise Air temperature modeling is crucial for a sustainable environment. In this study, a novel binary optimized machine learning model, the random vector functional link (RVFL) with the integration of Moth Flame Optimization Algorithm (MFO) and Water Cycle Optimization Algorithm (WCA) is examined to estimate the monthly and daily temperature time series of Rajshahi Climatic station in Bangladesh. Various combinations of temperature and precipitation were used to predict the temperature time series. The prediction ability of the novel binary optimized machine learning model (RVFL-WCAMFO) is compared with the single optimized machine learning models (RVFL-WCA and RVFL-MFO) and the standalone machine learning model (RVFL). Root mean square errors (RMSE), the mean absolute error (MAE), the Nash–Sutcliffe efficiency (NSE), and the determination coefficient (R2) statistical indexes were utilized to access the prediction ability of the selected models. The proposed binary optimized machine learning model (RVFL-WCAMFO) outperformed the other single optimized and standalone machine learning models in prediction of air temperature time series on both scales, i.e., daily and monthly scale. Cross-validation technique was applied to determine the best testing dataset and it was found that the M3 dataset provided more accurate results for the monthly scale, whereas the M1 dataset outperformed the other two datasets on the daily scale. On the monthly scale, periodicity input was also added to see the effect on prediction accuracy. It was found that periodicity input improved the prediction accuracy of the models. It was also found that precipitation-based inputs did not provided very accurate results in comparison to temperature-based inputs. The outcomes of the study recommend the use of RVFL-WCAMFO in air temperature modeling.

A hybrid neural network and optimization algorithm for forecasting and trend detection of Forex market indices

An Autoencoder (AE) is an independent feature extractor from data samples and a deep network can be obtained by stacking several AEs. This paper presents a novel hybrid stacked Autoencoder-based Deep Kernel-based Random Vector Functional Link Network (DKRVFLN-AE) for forecasting and trend analysis of Foreign Exchange (Forex) rates. The proposed model dispenses the random choices of weights and biases, unlike the stacked Deep Random Vector Functional Link Network-AE (DRVFLN-AE), using a wavelet kernel function with strong data fitting capability based on Mercer’s condition. A modified metaheuristic Water Cycle Algorithm is used to optimize the wavelet kernel parameters and provide DKRVFLN-AE a better generalization and learning capability, faster execution speed, lower storage space, and improved accuracy to traditional deep learning Random Vector Functional Link Network and Extreme Learning Machine models. Applications of this new approach to predict exchange rates and trend analysis on three foreign exchange markets provide successful results and validate its superiority over well-known approaches like Random Vector Functional Link Networks, Support Vector Machines, Naive-Bayes, Extreme Learning Machines, and Deep Random Vector Functional Link Network.

Factorization of broad expansion for broad learning system

The broad learning system (BLS) based on the random vector functional link neural network is a new versatile non-iterative neural network for rapidly selecting models. One of its fascinating features is the broad expansion, a dynamic adjustment strategy in the changing environment. However, the added enhancement nodes often suffer from information redundancy and loss, a failure in the broad expansion. Therefore, the factorization of enhancement nodes, which makes each code independent of others and achieves minimal redundancy and loss representation, is critical to successful broad expansion. Inspired by this idea, in this study, a factorial broad expansion (FBE) strategy is proposed to aid BLS (FBE-BLS) during broad expansion. The strategy is proposed as a component for reducing the information redundancy and loss in the classical BLS. This strategy is an adversarial framework to learn the factorial codes at the enhancement layer. The factorial and predictor networks adversarially learn factorial codes by chasing independence, minimizing predictability among enhancement nodes. In experiments, considering that few works focused on information redundancy and loss in the broad expansion of BLS, the FBE-BLS is compared with the classical BLS, and state-of-the-art methods are selected. The results of FBE-BLS on UCI and real-world image data are very promising, exhibiting higher accuracy during broad expansion.

Applications of machine learning in friction stir welding: Prediction of joint properties, real-time control and tool failure diagnosis

Machine learning (ML) methods have received immense attention as potential models for modeling different manufacturing systems. This paper presents a comprehensive review on the applications of ML methods in friction stir welding (FSW) field. Five main topics have been discussed: prediction of the joint properties, integration between ML and finite element methods, real-time control of FSW process, tool failure diagnosis, and incorporation between metaheuristic optimization techniques and ML methods. The common used ML methods such as multi-linear regression, K-nearest neighbor, random forest algorithm, Gaussian process regression, artificial neural network, support vector machine, radial basis function neural network, fuzzy system, adaptive neuro-fuzzy inference system, and random vector functional link are explained. Then, different statistical measures used to evaluate the performance of ML methods are presented. Finally, the applications of ML methods in FSW field are discussed. Important conclusions are drawn and future prospects are suggested.

Investigating White Matter Abnormalities Associated with Schizophrenia Using Deep Learning Model and Voxel-Based Morphometry

Schizophrenia (SCZ) is a devastating mental condition with significant negative consequences for patients, making correct and prompt diagnosis crucial. The purpose of this study is to use structural magnetic resonance image (MRI) to better classify individuals with SCZ from control normals (CN) and to locate a region of the brain that represents abnormalities associated with SCZ. Deep learning (DL), which is based on the nervous system, could be a very useful tool for doctors to accurately predict, diagnose, and treat SCZ. Gray Matter (GM), Cerebrospinal Fluid (CSF), and White Matter (WM) brain regions are extracted from 99 MRI images obtained from the open-source OpenNeuro database to demonstrate SCZ's regional relationship. In this paper, we use a pretrained ResNet-50 deep network to extract features from MRI images and an ensemble deep random vector functional link (edRVFL) network to classify those features. By examining the results obtained, the edRVFL deep model provides the highest classification accuracy of 96.5% with WM and is identified as the best-performing algorithm compared to the traditional algorithms. Furthermore, we examined the GM, WM, and CSF tissue volumes in CN subjects and SCZ patients using voxel-based morphometry (VBM), and the results show 1363 significant voxels, 6.90 T-value, and 6.21 Z-value in the WM region of SCZ patients. In SCZ patients, WM is most closely linked to structural alterations, as evidenced by VBM analysis and the DL model.

Modeling Potential Evapotranspiration by Improved Machine Learning Methods Using Limited Climatic Data

Modeling potential evapotranspiration (ET0) is an important issue for water resources planning and management projects involving droughts and flood hazards. Evapotranspiration, one of the main components of the hydrological cycle, is highly effective in drought monitoring. This study investigates the efficiency of two machine-learning methods, random vector functional link (RVFL) and relevance vector machine (RVM), improved with new metaheuristic algorithms, quantum-based avian navigation optimizer algorithm (QANA), and artificial hummingbird algorithm (AHA) in modeling ET0 using limited climatic data, minimum temperature, maximum temperature, and extraterrestrial radiation. The outcomes of the hybrid RVFL-AHA, RVFL-QANA, RVM-AHA, and RVM-QANA models compared with single RVFL and RVM models. Various input combinations and three data split scenarios were employed. The results revealed that the AHA and QANA considerably improved the efficiency of RVFL and RVM methods in modeling ET0. Considering the periodicity component and extraterrestrial radiation as inputs improved the prediction accuracy of the applied methods.

Fabric Wrinkle Objective Evaluation Model with Random Vector Function Link Based on Optimized Artificial Hummingbird Algorithm

ABSTRACT With the more and more amelioration of our quality of life, our needs for clothing have altered from having clothes to wearing good-looking, among which the wrinkle resistance of clothing fabric owns a giant effect on the beauty of clothing. Nowadays, artificial subjective evaluation is mainly used to evaluate the wrinkle grade of garment fabrics in the textile industry. This evaluation method owns the shortcoming of poor accuracy, being time-consuming and poor objectivity. For solving this problem, it is very important to put forward an objective evaluation model of fabric wrinkle grade. In this paper, we proposed a fabric wrinkle objective evaluation model with the optimized random vector functional link. The model applies DarkNet19 deep neural network to abstract the high-order visual features of the wrinkled surface image of the fabric, uses the improved artificial hummingbird optimization algorithm to ameliorate the import bias and weight of the random vector function link’s hidden layer, and uses L 2 , 1 norm regularization computes output weights for random vector function links. The relative tests reveal that the objective evaluation model of fabric wrinkles put forward in this paper has excellent performance.

Water distillation tower: Experimental investigation, economic assessment, and performance prediction using optimized machine-learning model

Here we present a new compact design of a vertical water distillation tower based on solar stills. The experimental setup consisted of a vertical tower with five water trays, supported by a metal duct and surrounded by a glass enclosure. The water yield and thermal, exergic, and economic features of the tower were investigated and analyzed. A new performance prediction hybrid model was also developed. A powerful artificial intelligence tool called the random vector functional link (RVFL) neural network was integrated with the Runge Kutta optimizer (RUN) to predict the water yield and temperature of the established tower. The model efficiency was compared to that of pure RVFL and an optimized RVFL model using a particle swarm optimizer (PSO). The drinkable water yield of the proposed design was 2.1 L/m2 (considering the tray area) and 5.3 L/m2 (considering the land use area); energy and exergy efficiencies were ∼31.7% and ∼3.3%, respectively. The cost of the produced drinkable water was approximately $0.013/L. The developed system provides considerable improvement compared with conventional designs of solar stills. The proposed RVFL–RUN model outperformed the pure RVFL and RVFL–PSO models for predicting system performance. The coefficients of determination between the experimental water productivity and water temperature and the predicted values, using the RVFL–RUN model, were 0.91 and 0.97, respectively.

Artificially intelligent differential diagnosis of enlarged lymph nodes with random vector functional link network plus.

Differential diagnosis of enlarged lymph nodes (ELNs) is essential for the treatment of related patients. Though multi-modal ultrasound including B-mode, Doppler ultrasound, elastographyand contrast-enhanced ultrasound (CEUS)can enhance diagnostic performance for ELNs, the scenario of having only single or dual modal data is often encountered. In this study, an artificially intelligent diagnosis model based on the learning using privileged information was proposed to aid in differential diagnosis of ELNs in the case of single or dual modal images. In our model, B-mode, or combined with another modality, was used as the standard information (SI) and other modalities were used as the privileged information (PI). The model was constructed through the combination of the SI and PI in the training stage. By learning from the training samples, a random vector functional link network with privileged information (RVFL+) was obtained, which was used to classify the testing samples of solely the SI. Results showed that the accuracy, precision and Youden's index of the RVFL+ model, using B-mode with elastography as the SI and CEUS as the PI, reached 78.4%, 92.4% and 54.9%, increased by 14.0%, 8.4% and 24.5% compared with the model using B-mode as the SI without the PI. The method based on the LUPI can improve the diagnostic performance for ELNs.

Teamwork Optimization with Deep Learning Based Fall Detection for營oT-Enabled Smart Healthcare System

The current advancement in cloud computing, Artificial Intelligence (AI), and the Internet of Things (IoT) transformed the traditional healthcare system into smart healthcare. Healthcare services could be enhanced by incorporating key techniques like AI and IoT. The convergence of AI and IoT provides distinct opportunities in the medical field. Fall is regarded as a primary cause of death or post-traumatic complication for the ageing population. Therefore, earlier detection of older person falls in smart homes is required to improve the survival rate of an individual or provide the necessary support. Lately, the emergence of IoT, AI, smartphones, wearables, and so on making it possible to design fall detection (FD) systems for smart home care. This article introduces a new Teamwork Optimization with Deep Learning based Fall Detection for IoT Enabled Smart Healthcare Systems (TWODL-FDSHS). The TWODL-FDSHS technique’s goal is to detect fall events for a smart healthcare system. Initially, the presented TWODL-FDSHS technique exploits IoT devices for the data collection process. Next, the TWODL-FDSHS technique applies the TWO with Capsule Network (CapsNet) model for feature extraction. At last, a deep random vector functional link network (DRVFLN) with an Adam optimizer is exploited for fall event detection. A wide range of simulations took place to exhibit the enhanced performance of the presented TWODL-FDSHS technique. The experimental outcomes stated the enhancements of the TWODL-FDSHS method over other models with increased accuracy of 98.30% on the URFD dataset.

An Elitist Artificial Electric Field Algorithm Based Random Vector Functional Link Network for Cryptocurrency Prices Forecasting

An Elitist Artificial Electric Field Algorithm Based Random Vector Functional Link Network for Cryptocurrency Prices Forecasting

Evolution-Driven Randomized Graph Convolutional Networks

Randomized neural networks (NNs), such as random vector functional link (RVFL) and extreme learning machine (ELM), have been widely applied in various classification problems owing to their computational efficiency and universal approximation capability. However, such approaches are designed for regular Euclidean data and lack the ability to generalize to complex structured data. Moreover, their randomly generated parameters often lead to a suboptimal decision boundary with a growing requirement of hidden neurons. In this article, we first propose a plain framework, termed randomized graph convolutional networks (RGCNs), to generalize the classic randomized NNs to the non-Euclidean domain. Then, a hybrid framework called evolution-driven RGCN (EvoRGCN) is presented by using adaptive differential evolution with novelty search strategy to seek the globally optimal graph embedding for the plain RGCN. Finally, we recast the classic ELM and RVFL under the proposed frameworks, resulting in four novel semi-supervised models, including the plain models [i.e., graph convolutional extreme learning machines (GCELMs) and graph convolutional RVFL (GCRVFL)] and the optimized models (i.e., O-GCELM and O-GCRVFL). We show that our approaches are the natural generalization of the traditional randomized NNs in the non-Euclidean domain. Furthermore, our approaches not only retain the advantages of the classic approaches but also enable them to handle graph data. We compare our approaches against many existing methods across regular datasets and graph benchmarks, demonstrating that the proposed approaches dramatically outperform the compared methods with better generalization ability and robustness. Particularly, we quantitatively show the performance ranking of different randomized NNs, i.e., O-GCRVFL <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$&gt; $ </tex-math></inline-formula> O-GCELM <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\approx $ </tex-math></inline-formula> GCRVFL <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$&gt; $ </tex-math></inline-formula> GCELM <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\approx $ </tex-math></inline-formula> RVFL <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$&gt; $ </tex-math></inline-formula> ELM.

A Bayesian Approach To Distributed Anomaly Detection In Edge AI Networks

The challenge of anomaly detection is to obtain an accurate understanding of expected behaviour which is intensified when the data are distributed heterogeneously. Transmitting raw data to a central site incurs high communication overhead and raises privacy issues. The concept of Edge AI allows computation to be performed at the edge site allowing for quick decision making in mission critical scenarios such as self-driving cars. A model is learnt locally and its parameters are transmitted and aggregated. However, existing methods of aggregation do not account for variance and heterogeneous distribution of data. They also do not consider edge constraints such as limited computational, memory and communication capabilities of edge devices. In this work, a fully Bayesian approach is employed by means of a Bayesian Random Vector Functional Link AutoEncoder being incorporated with Expectation Propagation for distributed training. Our anomaly detection system operates without any transmission of raw data, is robust under inhomogeneous network densities and under uneven and biased data distributions. It allows for asynchronous updates to converge in a few iterations and is a relatively simple neural network addressing edge constraints without compromising on performance as compared to existing more complex models.

An Effective Online Sequential Stochastic Configuration Algorithm for Neural Networks

Random Vector Functional-link (RVFL) networks, as a class of random learner models, have received careful attention from the neural network research community due to their advantages in obtaining fast learning algorithms and models, in which the hidden layer parameters are randomly generated and remain fixed during the training phase. However, its universal approximation ability may not be guaranteed if the random parameters are not properly selected in an appropriate range. Moreover, the resulting random learner’s generalization performance may seriously deteriorate once the RVFL network’s structure is not well-designed. Stochastic configuration (SC) algorithm, which incrementally constructs a universal approximator by obtaining random hidden parameters under a specified supervisory mechanism, instead of fixing the selection scope in advance and without any reference to training information, can effectively circumvent these awkward issues caused by randomness. This paper extends the SC algorithm to an online sequential version, termed as an OSSC algorithm, by means of recursive least square (RLS) technique, aiming to copy with modeling tasks where training observations are sequentially provided. Compared to the online sequential learning of RVFL networks (OS-RVFL in short), our proposed OSSC algorithm can avoid the awkward setting of certain unreasonable range for the random parameters, and can also successfully build a random learner with preferable learning and generalization capabilities. The experimental study has shown the effectiveness and advantages of our OSSC algorithm.

Error-output recurrent multi-layer Kernel Reservoir Network for electricity load time series forecasting

Electricity is one of the most consumed commodities in the modern world. Electricity load prediction models are used to plan distribution operations to balance the equilibrium of demand and supply. This necessity has increased the number of recent research works. They employed several learning algorithms, such as support vector regression, to predict demands. However, these algorithms have high computational cost and too many user-defined parameters that directly impact their performance. Recently, randomization-based learning algorithms have been widely tested because they performed well at a lower cost. However, still, there was a main drawback: uncertainty in approximation and learning. This work employed a kernel trick to solve the uncertainty problem. A kernel with reservoir-state layers was used to solve the problem. The kernel reservoir-state layers from the echo state network not only transformed features into high-dimensional space, but also enhanced the forecasting ability by learning temporal information. Additionally, the proposed model also had a multi-step prediction ability that used previous forecasting errors to update the output weights in the current step to prevent an accumulated error problem. We compared our proposed model with single-layer and multi-layer variants of Extreme Learning Machine, Echo State Network, and Random Vector Functional Link on ten electrical load data sets. The proposed model showed the best performance on 9/10 data sets in terms of Mean Square Error or Symmetric Mean Absolute Percentage Error. These findings implied that the proposed algorithm was superior in forecasting long-term electricity load.

Spatio Prediction of Soil Capability Modeled with Modified RVFL Using Aptenodytes Forsteri Optimization and Digital Soil Assessment Technique

To meet the needs of Egypt’s rising population, more land must be cultivated. Land evaluation is vital to achieving sustainable agricultural production. To determine the soil capability in the northeast Nile Delta region of Egypt, the present study introduces a new form of integration between the Agriculture Land Evaluation System (ALES Arid) model and the machine learning (ML) approach. The soil capability indicators required for the ALES Arid model were determined for the 47 collected soil profiles covering the study area. These indicators include soil pH, soil salinity, the sodium adsorption ratio (SAR), the exchangeable sodium percentage (ESP), the organic matter (OM) content, the calcium carbonate (CaCO3) content, the gypsum content, the clay percentage, and the slope. The ALES Arid model was run using these indicators, and soil capability indexes were obtained. Using GIS, these indexes helped to classify the study area into four capability classes, ranging from good to very poor soils. To predict the soil capability, three machine learning algorithms named traditional RVFL, sine cosine algorithm (SCA), and AFO were also applied to the same soil criteria. The developed ML method aims to enhance the prediction of soil capability. This method depends on improving the performance of Random Vector Functional Link (RVFL) using an optimization technique named Aptenodytes Forsteri Optimization (AFO). The operators of AFO were used to determine the best parameters of RVFL since traditional RVFL is sensitive to parameters. To assess the performance of the developed AFO-RVFL method, a set of real collected data was used. The experimental results illustrate the high efficacy of AFO-RVFL in the spatial prediction of soil capability. The correlations found in this study are critical for understanding the overall techniques for predicting soil capability.

Graph ensemble deep random vector functional link network for traffic forecasting

Traffic forecasting is crucial to achieving a smart city as it facilitates public transportation management, autonomous driving, and the resource relocation of the sharing economy. Traffic forecasting belongs to the challenging spatiotemporal forecasting task, which is highly demanding because of the complicated geospatial correlation between traffic nodes, inconsistent and highly non-linear temporal patterns due to various events, and sporadic traffic accidents. Previous graph neural network (GNN) models built for transportation forecasting feature the sophisticated structure and heavy computation cost as they combine the deep neural network and graph machine learning to capture the spatiotemporal dynamics for the whole transportation network. However, it may be more practical for practitioners to perform node-wise forecasting for specific nodes of interest rather than network-wise forecasting. To mitigate the gaps mentioned above, we propose a novel graph ensemble deep random vector functional link network (GEdRVFL) to forecast the future traffic volume by combining the well-performing ensemble deep random vector functional link (EdRVFL) with the graph convolution layer for a specific node and realize the node-wise traffic forecasting. After a comprehensive comparison with the state-of-the-art models, our model beats the others in four out of five cases measured by mean absolute scaled error.