Neural networks(NNs),as one of the most robust and efficient machine learning methods,have been commonly used in solving several problems.However,choosing proper hyperparameters(e.g.the numbers of layers and neurons i...Neural networks(NNs),as one of the most robust and efficient machine learning methods,have been commonly used in solving several problems.However,choosing proper hyperparameters(e.g.the numbers of layers and neurons in each layer)has a significant influence on the accuracy of these methods.Therefore,a considerable number of studies have been carried out to optimize the NN hyperpaxameters.In this study,the genetic algorithm is applied to NN to find the optimal hyperpaxameters.Thus,the deep energy method,which contains a deep neural network,is applied first on a Timoshenko beam and a plate with a hole.Subsequently,the numbers of hidden layers,integration points,and neurons in each layer are optimized to reach the highest accuracy to predict the stress distribution through these structures.Thus,applying the proper optimization method on NN leads to significant increase in the NN prediction accuracy after conducting the optimization in various examples.展开更多
The selection of hyperparameters in regularized least squares plays an important role in large-scale system identification. The traditional methods for selecting hyperparameters are based on experience or marginal lik...The selection of hyperparameters in regularized least squares plays an important role in large-scale system identification. The traditional methods for selecting hyperparameters are based on experience or marginal likelihood maximization method, which are inaccurate or computationally expensive. In this paper, two posterior methods are proposed to select hyperparameters based on different prior knowledge (constraints), which can obtain the optimal hyperparameters using the optimization theory. Moreover, we also give the theoretical optimal constraints, and verify its effectiveness. Numerical simulation shows that the hyperparameters and parameter vector estimate obtained by the proposed methods are the optimal ones.展开更多
Analyzing big data, especially medical data, helps to provide good health care to patients and face the risks of death. The COVID-19 pandemic has had a significant impact on public health worldwide, emphasizing the ne...Analyzing big data, especially medical data, helps to provide good health care to patients and face the risks of death. The COVID-19 pandemic has had a significant impact on public health worldwide, emphasizing the need for effective risk prediction models. Machine learning (ML) techniques have shown promise in analyzing complex data patterns and predicting disease outcomes. The accuracy of these techniques is greatly affected by changing their parameters. Hyperparameter optimization plays a crucial role in improving model performance. In this work, the Particle Swarm Optimization (PSO) algorithm was used to effectively search the hyperparameter space and improve the predictive power of the machine learning models by identifying the optimal hyperparameters that can provide the highest accuracy. A dataset with a variety of clinical and epidemiological characteristics linked to COVID-19 cases was used in this study. Various machine learning models, including Random Forests, Decision Trees, Support Vector Machines, and Neural Networks, were utilized to capture the complex relationships present in the data. To evaluate the predictive performance of the models, the accuracy metric was employed. The experimental findings showed that the suggested method of estimating COVID-19 risk is effective. When compared to baseline models, the optimized machine learning models performed better and produced better results.展开更多
Background:The existence of doublets in single-cell RNA sequencing(scRNA-seq)data poses a great challenge in downstream data analysis.Computational doublet-detection methods have been developed to remove doublets from...Background:The existence of doublets in single-cell RNA sequencing(scRNA-seq)data poses a great challenge in downstream data analysis.Computational doublet-detection methods have been developed to remove doublets from scRNA-seq data.Yet,the default hyperparameter settings of those methods may not provide optimal performance.Methods:We propose a strategy to tune hyperparameters for a cutting-edge doublet-detection method.We utilize a full factorial design to explore the relationship between hyperparameters and detection accuracy on 16 real scRNA-seq datasets.The optimal hyperparameters are obtained by a response surface model and convex optimization.Results:We show that the optimal hyperparameters provide top performance across scRNA-seq datasets under various biological conditions.Our tuning strategy can be applied to other computational doublet-detection methods.It also offers insights into hyperparameter tuning for broader computational methods in scRNA-seq data analysis.Conclusions:The hyperparameter configuration significantly impacts the performance of computational doublet-detection methods.Our study is the first attempt to systematically explore the optimal hyperparameters under various biological conditions and optimization objectives.Our study provides much-needed guidance for hyperparameter tuning in computational doublet-detection methods.展开更多
To predict stall and surge in advance that make the aero-engine compressor operatesafely,a stall prediction model based on deep learning theory is established in the current study.The Long Short-Term Memory(LSTM)origi...To predict stall and surge in advance that make the aero-engine compressor operatesafely,a stall prediction model based on deep learning theory is established in the current study.The Long Short-Term Memory(LSTM)originating from the recurrent neural network is used,and a set of measured dynamic pressure datasets including the stall process is used to learn whatdetermines the weight of neural network nodes.Subsequently,the structure and function hyperpa-rameters in the model are deeply optimized,and a set of measured pressure data is used to verify theprediction effects of the model.On this basis of the above good predictive capability,stall in low-and high-speed compressor are predicted by using the established model.When a period of non-stallpressure data is used as input in the model,the model can quickly complete the prediction of sub-sequent time series data through the self-learning and prediction mechanism.Comparison with thereal-time measured pressure data demonstrates that the starting point of the predicted stall is basi-cally the same as that of the measured stall,and the stall can be predicted more than 1 s in advanceso that the occurrence of stall can be avoided.The model of stall prediction in the current study canmake up for the uncertainty of threshold selection of the existing stall warning methods based onmeasured data signal processing.It has a great application potential to predict the stall occurrenceof aero-engine compressor in advance and avoid the accidents.展开更多
To overcome the challenges associated with predicting gas extraction performance and mitigating the gradual decline in extraction volume,which adversely impacts gas utilization efficiency in mines,a gas extraction pur...To overcome the challenges associated with predicting gas extraction performance and mitigating the gradual decline in extraction volume,which adversely impacts gas utilization efficiency in mines,a gas extraction pure volume prediction model was developed using Support Vector Regression(SVR)and Random Forest(RF),with hyperparameters fine-tuned via the Genetic Algorithm(GA).Building upon this,an adaptive control model for gas extraction negative pressure was formulated to maximize the extracted gas volume within the pipeline network,followed by field validation experiments.Experimental results indicate that the GA-SVR model surpasses comparable models in terms of mean absolute error,root mean square error,and mean absolute percentage error.In the extraction process of bedding boreholes,the influence of negative pressure on gas extraction concentration diminishes over time,yet it remains a critical factor in determining the extracted pure volume.In contrast,throughout the entire extraction period of cross-layer boreholes,both extracted pure volume and concentration exhibit pronounced sensitivity to fluctuations in extraction negative pressure.Field experiments demonstrated that the adaptive controlmodel enhanced the average extracted gas volume by 5.08% in the experimental borehole group compared to the control group during the later extraction stage,with a more pronounced increase of 7.15% in the first 15 days.The research findings offer essential technical support for the efficient utilization and long-term sustainable development of mine gas resources.The research findings offer essential technical support for gas disaster mitigation and the sustained,efficient utilization of mine gas.展开更多
Traffic forecasting with high precision aids Intelligent Transport Systems(ITS)in formulating and optimizing traffic management strategies.The algorithms used for tuning the hyperparameters of the deep learning models...Traffic forecasting with high precision aids Intelligent Transport Systems(ITS)in formulating and optimizing traffic management strategies.The algorithms used for tuning the hyperparameters of the deep learning models often have accurate results at the expense of high computational complexity.To address this problem,this paper uses the Tree-structured Parzen Estimator(TPE)to tune the hyperparameters of the Long Short-term Memory(LSTM)deep learning framework.The Tree-structured Parzen Estimator(TPE)uses a probabilistic approach with an adaptive searching mechanism by classifying the objective function values into good and bad samples.This ensures fast convergence in tuning the hyperparameter values in the deep learning model for performing prediction while still maintaining a certain degree of accuracy.It also overcomes the problem of converging to local optima and avoids timeconsuming random search and,therefore,avoids high computational complexity in prediction accuracy.The proposed scheme first performs data smoothing and normalization on the input data,which is then fed to the input of the TPE for tuning the hyperparameters.The traffic data is then input to the LSTM model with tuned parameters to perform the traffic prediction.The three optimizers:Adaptive Moment Estimation(Adam),Root Mean Square Propagation(RMSProp),and Stochastic Gradient Descend with Momentum(SGDM)are also evaluated for accuracy prediction and the best optimizer is then chosen for final traffic prediction in TPE-LSTM model.Simulation results verify the effectiveness of the proposed model in terms of accuracy of prediction over the benchmark schemes.展开更多
In radiology,magnetic resonance imaging(MRI)is an essential diagnostic tool that provides detailed images of a patient’s anatomical and physiological structures.MRI is particularly effective for detecting soft tissue...In radiology,magnetic resonance imaging(MRI)is an essential diagnostic tool that provides detailed images of a patient’s anatomical and physiological structures.MRI is particularly effective for detecting soft tissue anomalies.Traditionally,radiologists manually interpret these images,which can be labor-intensive and time-consuming due to the vast amount of data.To address this challenge,machine learning,and deep learning approaches can be utilized to improve the accuracy and efficiency of anomaly detection in MRI scans.This manuscript presents the use of the Deep AlexNet50 model for MRI classification with discriminative learning methods.There are three stages for learning;in the first stage,the whole dataset is used to learn the features.In the second stage,some layers of AlexNet50 are frozen with an augmented dataset,and in the third stage,AlexNet50 with an augmented dataset with the augmented dataset.This method used three publicly available MRI classification datasets:Harvard whole brain atlas(HWBA-dataset),the School of Biomedical Engineering of Southern Medical University(SMU-dataset),and The National Institute of Neuroscience and Hospitals brain MRI dataset(NINS-dataset)for analysis.Various hyperparameter optimizers like Adam,stochastic gradient descent(SGD),Root mean square propagation(RMS prop),Adamax,and AdamW have been used to compare the performance of the learning process.HWBA-dataset registers maximum classification performance.We evaluated the performance of the proposed classification model using several quantitative metrics,achieving an average accuracy of 98%.展开更多
Fire can cause significant damage to the environment,economy,and human lives.If fire can be detected early,the damage can be minimized.Advances in technology,particularly in computer vision powered by deep learning,ha...Fire can cause significant damage to the environment,economy,and human lives.If fire can be detected early,the damage can be minimized.Advances in technology,particularly in computer vision powered by deep learning,have enabled automated fire detection in images and videos.Several deep learning models have been developed for object detection,including applications in fire and smoke detection.This study focuses on optimizing the training hyperparameters of YOLOv8 andYOLOv10models usingBayesianTuning(BT).Experimental results on the large-scale D-Fire dataset demonstrate that this approach enhances detection performance.Specifically,the proposed approach improves the mean average precision at an Intersection over Union(IoU)threshold of 0.5(mAP50)of the YOLOv8s,YOLOv10s,YOLOv8l,and YOLOv10lmodels by 0.26,0.21,0.84,and 0.63,respectively,compared tomodels trainedwith the default hyperparameters.The performance gains are more pronounced in larger models,YOLOv8l and YOLOv10l,than in their smaller counterparts,YOLOv8s and YOLOv10s.Furthermore,YOLOv8 models consistently outperform YOLOv10,with mAP50 improvements of 0.26 for YOLOv8s over YOLOv10s and 0.65 for YOLOv8l over YOLOv10l when trained with BT.These results establish YOLOv8 as the preferred model for fire detection applications where detection performance is prioritized.展开更多
With the rapid adoption of artificial intelligence(AI)in domains such as power,transportation,and finance,the number of machine learning and deep learning models has grown exponentially.However,challenges such as dela...With the rapid adoption of artificial intelligence(AI)in domains such as power,transportation,and finance,the number of machine learning and deep learning models has grown exponentially.However,challenges such as delayed retraining,inconsistent version management,insufficient drift monitoring,and limited data security still hinder efficient and reliable model operations.To address these issues,this paper proposes the Intelligent Model Lifecycle Management Algorithm(IMLMA).The algorithm employs a dual-trigger mechanism based on both data volume thresholds and time intervals to automate retraining,and applies Bayesian optimization for adaptive hyperparameter tuning to improve performance.A multi-metric replacement strategy,incorporating MSE,MAE,and R2,ensures that new models replace existing ones only when performance improvements are guaranteed.A versioning and traceability database supports comparison and visualization,while real-time monitoring with stability analysis enables early warnings of latency and drift.Finally,hash-based integrity checks secure both model files and datasets.Experimental validation in a power metering operation scenario demonstrates that IMLMA reduces model update delays,enhances predictive accuracy and stability,and maintains low latency under high concurrency.This work provides a practical,reusable,and scalable solution for intelligent model lifecycle management,with broad applicability to complex systems such as smart grids.展开更多
The 91 measured values of the development height of the water-conducting fracture zone(WCFZ)in deep and thick coal seam mining faces under thick loose layer conditions were collected.Five key characteristic variables ...The 91 measured values of the development height of the water-conducting fracture zone(WCFZ)in deep and thick coal seam mining faces under thick loose layer conditions were collected.Five key characteristic variables influencing the WCFZ height were identified.After removing outliers from the dataset,a Random Forest(RF)regression model optimized by the Sparrow Search Algorithm(SSA)was constructed.The hyperparameters of the RF model were iteratively optimized by minimizing the Out-of-Bag(OOB)error,resulting in the rapid deter-mination of optimal parameters.Specifically,the SSA-RF model achieved an OOB error of 0.148,with 20 de-cision trees,a maximum depth of 8,a minimum split sample size of 2,and a minimum leaf node sample size of 1.Cross-validation experiments were performed using the trained optimal model and compared against other prediction methods.The results showed that the mining height had the most significant correlation with the development height of the WCFZ.The SSA-RF model outperformed all other models,with R2 values exceeding 0.9 across the training,validation,and test datasets.Compared to other models,the SSA-RF model demonstrates a simpler structure,stronger fitting capacity,higher predictive accuracy,and superior stability and generaliza-tion ability.It also exhibits the smallest variation in relative error across datasets,indicating excellent adapt-ability to different data conditions.Furthermore,a numerical model was developed using the hydrogeological data from the 1305 working face at Wanfukou Coal Mine,Shandong Province,China,to simulate the dynamic development of the WCFZ during mining.The SSA-RF model predicted the WCFZ height to be 69.7 m,closely aligning with the PFC2D simulation result of 65 m,with an error of less than 5%.Compared to traditional methods and numerical simulations,the SSA-RF model provides more accurate predictions,showing only a 7.23% deviation from the PFC2D simulation,while traditional empirical formulas yield deviations as large as 19.97%.These results demonstrate the SSA-RF model’s superior predictive capability,reinforcing its reliability and engineering applicability for real-world mining operations.This model holds significant potential for enhancing mining safety and optimizing planning processes,offering a more accurate and efficient approach for WCFZ height prediction.展开更多
Sparse identification of nonlinear dynamics(SINDy)has made significant progress in data-driven dynamics modeling.However,determining appropriate hyperparameters and addressing the time-consuming symbolic regression pr...Sparse identification of nonlinear dynamics(SINDy)has made significant progress in data-driven dynamics modeling.However,determining appropriate hyperparameters and addressing the time-consuming symbolic regression process remain substantial challenges.This study proposes the adaptive backward stepwise selection of fast SINDy(ABSS-FSINDy),which integrates statistical learning-based estimation and technical advancements to significantly reduce simulation time.This approach not only provides insights into the conditions under which SINDy performs optimally but also highlights potential failure points,particularly in the context of backward stepwise selection(BSS).By decoding predefined features into textual expressions,ABSS-FSINDy significantly reduces the simulation time compared with conventional symbolic regression methods.We validate the proposed method through a series of numerical experiments involving both planar/spatial dynamics and high-dimensional chaotic systems,including Lotka-Volterra,hyperchaotic Rossler,coupled Lorenz,and Lorenz 96 benchmark systems.The experimental results demonstrate that ABSS-FSINDy autonomously determines optimal hyperparameters within the SINDy framework,overcoming the curse of dimensionality in high-dimensional simulations.This improvement is substantial across both lowand high-dimensional systems,yielding efficiency gains of one to three orders of magnitude.For instance,in a 20D dynamical system,the simulation time is reduced from 107.63 s to just 0.093 s,resulting in a 3-order-of-magnitude improvement in simulation efficiency.This advancement broadens the applicability of SINDy for the identification and reconstruction of high-dimensional dynamical systems.展开更多
Recently,deep learning(DL)became one of the essential tools in bioinformatics.A modified convolutional neural network(CNN)is employed in this paper for building an integratedmodel for deoxyribonucleic acid(DNA)classif...Recently,deep learning(DL)became one of the essential tools in bioinformatics.A modified convolutional neural network(CNN)is employed in this paper for building an integratedmodel for deoxyribonucleic acid(DNA)classification.In any CNN model,convolutional layers are used to extract features followed by max-pooling layers to reduce the dimensionality of features.A novel method based on downsampling and CNNs is introduced for feature reduction.The downsampling is an improved form of the existing pooling layer to obtain better classification accuracy.The two-dimensional discrete transform(2D DT)and two-dimensional random projection(2D RP)methods are applied for downsampling.They convert the high-dimensional data to low-dimensional data and transform the data to the most significant feature vectors.However,there are parameters which directly affect how a CNN model is trained.In this paper,some issues concerned with the training of CNNs have been handled.The CNNs are examined by changing some hyperparameters such as the learning rate,size of minibatch,and the number of epochs.Training and assessment of the performance of CNNs are carried out on 16S rRNA bacterial sequences.Simulation results indicate that the utilization of a CNN based on wavelet subsampling yields the best trade-off between processing time and accuracy with a learning rate equal to 0.0001,a size of minibatch equal to 64,and a number of epochs equal to 20.展开更多
Since 2019,the coronavirus disease-19(COVID-19)has been spreading rapidly worldwide,posing an unignorable threat to the global economy and human health.It is a disease caused by severe acute respiratory syndrome coron...Since 2019,the coronavirus disease-19(COVID-19)has been spreading rapidly worldwide,posing an unignorable threat to the global economy and human health.It is a disease caused by severe acute respiratory syndrome coronavirus 2,a single-stranded RNA virus of the genus Betacoronavirus.This virus is highly infectious and relies on its angiotensin-converting enzyme 2-receptor to enter cells.With the increase in the number of confirmed COVID-19 diagnoses,the difficulty of diagnosis due to the lack of global healthcare resources becomes increasingly apparent.Deep learning-based computer-aided diagnosis models with high generalisability can effectively alleviate this pressure.Hyperparameter tuning is essential in training such models and significantly impacts their final performance and training speed.However,traditional hyperparameter tuning methods are usually time-consuming and unstable.To solve this issue,we introduce Particle Swarm Optimisation to build a PSO-guided Self-Tuning Convolution Neural Network(PSTCNN),allowing the model to tune hyperparameters automatically.Therefore,the proposed approach can reduce human involvement.Also,the optimisation algorithm can select the combination of hyperparameters in a targeted manner,thus stably achieving a solution closer to the global optimum.Experimentally,the PSTCNN can obtain quite excellent results,with a sensitivity of 93.65%±1.86%,a specificity of 94.32%±2.07%,a precision of 94.30%±2.04%,an accuracy of 93.99%±1.78%,an F1-score of 93.97%±1.78%,Matthews Correlation Coefficient of 87.99%±3.56%,and Fowlkes-Mallows Index of 93.97%±1.78%.Our experiments demonstrate that compared to traditional methods,hyperparameter tuning of the model using an optimisation algorithm is faster and more effective.展开更多
Abnormalities of the gastrointestinal tract are widespread worldwide today.Generally,an effective way to diagnose these life-threatening diseases is based on endoscopy,which comprises a vast number of images.However,t...Abnormalities of the gastrointestinal tract are widespread worldwide today.Generally,an effective way to diagnose these life-threatening diseases is based on endoscopy,which comprises a vast number of images.However,the main challenge in this area is that the process is time-consuming and fatiguing for a gastroenterologist to examine every image in the set.Thus,this led to the rise of studies on designingAI-based systems to assist physicians in the diagnosis.In several medical imaging tasks,deep learning methods,especially convolutional neural networks(CNNs),have contributed to the stateof-the-art outcomes,where the complicated nonlinear relation between target classes and data can be learned and not limit to hand-crafted features.On the other hand,hyperparameters are commonly set manually,which may take a long time and leave the risk of non-optimal hyperparameters for classification.An effective tool for tuning optimal hyperparameters of deep CNNis Bayesian optimization.However,due to the complexity of the CNN,the network can be regarded as a black-box model where the information stored within it is hard to interpret.Hence,Explainable Artificial Intelligence(XAI)techniques are applied to overcome this issue by interpreting the decisions of the CNNs in such wise the physicians can trust.To play an essential role in real-time medical diagnosis,CNN-based models need to be accurate and interpretable,while the uncertainty must be handled.Therefore,a novel method comprising of three phases is proposed to classify these life-threatening diseases.At first,hyperparameter tuning is performed using Bayesian optimization for two state-of-the-art deep CNNs,and then Darknet53 and InceptionV3 features are extracted from these fine-tunned models.Secondly,XAI techniques are used to interpret which part of the images CNN takes for feature extraction.At last,the features are fused,and uncertainties are handled by selecting entropybased features.The experimental results show that the proposed method outperforms existing methods by achieving an accuracy of 97%based on a Bayesian optimized Support Vector Machine classifier.展开更多
With the advancement of artificial intelligence,traffic forecasting is gaining more and more interest in optimizing route planning and enhancing service quality.Traffic volume is an influential parameter for planning ...With the advancement of artificial intelligence,traffic forecasting is gaining more and more interest in optimizing route planning and enhancing service quality.Traffic volume is an influential parameter for planning and operating traffic structures.This study proposed an improved ensemble-based deep learning method to solve traffic volume prediction problems.A set of optimal hyperparameters is also applied for the suggested approach to improve the performance of the learning process.The fusion of these methodologies aims to harness ensemble empirical mode decomposition’s capacity to discern complex traffic patterns and long short-term memory’s proficiency in learning temporal relationships.Firstly,a dataset for automatic vehicle identification is obtained and utilized in the preprocessing stage of the ensemble empirical mode decomposition model.The second aspect involves predicting traffic volume using the long short-term memory algorithm.Next,the study employs a trial-and-error approach to select a set of optimal hyperparameters,including the lookback window,the number of neurons in the hidden layers,and the gradient descent optimization.Finally,the fusion of the obtained results leads to a final traffic volume prediction.The experimental results show that the proposed method outperforms other benchmarks regarding various evaluation measures,including mean absolute error,root mean squared error,mean absolute percentage error,and R-squared.The achieved R-squared value reaches an impressive 98%,while the other evaluation indices surpass the competing.These findings highlight the accuracy of traffic pattern prediction.Consequently,this offers promising prospects for enhancing transportation management systems and urban infrastructure planning.展开更多
Deep learning techniques,particularly convolutional neural networks(CNNs),have exhibited remarkable performance in solving visionrelated problems,especially in unpredictable,dynamic,and challenging environments.In aut...Deep learning techniques,particularly convolutional neural networks(CNNs),have exhibited remarkable performance in solving visionrelated problems,especially in unpredictable,dynamic,and challenging environments.In autonomous vehicles,imitation-learning-based steering angle prediction is viable due to the visual imagery comprehension of CNNs.In this regard,globally,researchers are currently focusing on the architectural design and optimization of the hyperparameters of CNNs to achieve the best results.Literature has proven the superiority of metaheuristic algorithms over the manual-tuning of CNNs.However,to the best of our knowledge,these techniques are yet to be applied to address the problem of imitationlearning-based steering angle prediction.Thus,in this study,we examine the application of the bat algorithm and particle swarm optimization algorithm for the optimization of the CNN model and its hyperparameters,which are employed to solve the steering angle prediction problem.To validate the performance of each hyperparameters’set and architectural parameters’set,we utilized the Udacity steering angle dataset and obtained the best results at the following hyperparameter set:optimizer,Adagrad;learning rate,0.0052;and nonlinear activation function,exponential linear unit.As per our findings,we determined that the deep learning models show better results but require more training epochs and time as compared to shallower ones.Results show the superiority of our approach in optimizing CNNs through metaheuristic algorithms as compared with the manual-tuning approach.Infield testing was also performed using the model trained with the optimal architecture,which we developed using our approach.展开更多
Hydrological models are developed to simulate river flows over a watershed for many practical applications in the field of water resource management. The present paper compares the performance of two recurrent neural ...Hydrological models are developed to simulate river flows over a watershed for many practical applications in the field of water resource management. The present paper compares the performance of two recurrent neural networks for rainfall-runoff modeling in the Zou River basin at Atchérigbé outlet. To this end, we used daily precipitation data over the period 1988-2010 as input of the models, such as the Long Short-Term Memory (LSTM) and Recurrent Gate Networks (GRU) to simulate river discharge in the study area. The investigated models give good results in calibration (R2 = 0.888, NSE = 0.886, and RMSE = 0.42 for LSTM;R2 = 0.9, NSE = 0.9 and RMSE = 0.397 for GRU) and in validation (R2 = 0.865, NSE = 0.851, and RMSE = 0.329 for LSTM;R2 = 0.9, NSE = 0.865 and RMSE = 0.301 for GRU). This good performance of LSTM and GRU models confirms the importance of models based on machine learning in modeling hydrological phenomena for better decision-making.展开更多
Potato late blight and early blight are common hazards to the long-term production of potatoes, impacting many farmers around the world, particularly in Africa. Early detection and treatment of the potato blight disea...Potato late blight and early blight are common hazards to the long-term production of potatoes, impacting many farmers around the world, particularly in Africa. Early detection and treatment of the potato blight disease are critical for promoting healthy potato plant growth and ensuring adequate supply and food security for the fast-growing population. As a result, machine-driven disease detection systems may be able to overcome the constraints of traditional leaf disease diagnosis procedures, which are generally time-consuming, inaccurate, and costly. Convolutional Neural Networks (CNNs) have been shown to be effective in a variety of agricultural applications. CNNs have been shown to be helpful in detecting disease in plants because of their capacity to analyze vast volumes of data quickly and reliably. However, the method hasn’t been widely used in the detection of potato late blight and early blight diseases, which reduce yields significantly. The goal of this study was to compare six cutting-edge CNN architectural models, taking into account transfer learning for training and four hyperparameters. The CNN architectures evaluated were AlexNet, GoogleNet, SqueezeNet, DenseNet121, EfficientNet b7, and VGG19. Likewise, the hyperparameters analyzed were the number of epochs, the batch size, the optimizer, and the learning rate. An open-source dataset containing 4082 images was used. The DenseNet121 architecture with a batch of 32 and a Stochastic Gradient Descent (SGD) optimizer with a learning rate of 0.01 produced the best performance, with an accuracy of 98.34% and a 97.37% f1-score. The DenseNet121 model was shown to be useful in developing computer vision systems that aid farmers in improving their disease management systems for potato cultivation.展开更多
Hyperparameters are important for machine learning algorithms since they directly control the behaviors of training algorithms and have a significant effect on the performance of machine learning models. Several techn...Hyperparameters are important for machine learning algorithms since they directly control the behaviors of training algorithms and have a significant effect on the performance of machine learning models. Several techniques have been developed and successfully applied for certain application domains. However, this work demands professional knowledge and expert experience. And sometimes it has to resort to the brute-force search.Therefore, if an efficient hyperparameter optimization algorithm can be developed to optimize any given machine learning method, it will greatly improve the efficiency of machine learning. In this paper, we consider building the relationship between the performance of the machine learning models and their hyperparameters by Gaussian processes. In this way, the hyperparameter tuning problem can be abstracted as an optimization problem and Bayesian optimization is used to solve the problem. Bayesian optimization is based on the Bayesian theorem. It sets a prior over the optimization function and gathers the information from the previous sample to update the posterior of the optimization function. A utility function selects the next sample point to maximize the optimization function.Several experiments were conducted on standard test datasets. Experiment results show that the proposed method can find the best hyperparameters for the widely used machine learning models, such as the random forest algorithm and the neural networks, even multi-grained cascade forest under the consideration of time cost.展开更多
文摘Neural networks(NNs),as one of the most robust and efficient machine learning methods,have been commonly used in solving several problems.However,choosing proper hyperparameters(e.g.the numbers of layers and neurons in each layer)has a significant influence on the accuracy of these methods.Therefore,a considerable number of studies have been carried out to optimize the NN hyperpaxameters.In this study,the genetic algorithm is applied to NN to find the optimal hyperpaxameters.Thus,the deep energy method,which contains a deep neural network,is applied first on a Timoshenko beam and a plate with a hole.Subsequently,the numbers of hidden layers,integration points,and neurons in each layer are optimized to reach the highest accuracy to predict the stress distribution through these structures.Thus,applying the proper optimization method on NN leads to significant increase in the NN prediction accuracy after conducting the optimization in various examples.
文摘The selection of hyperparameters in regularized least squares plays an important role in large-scale system identification. The traditional methods for selecting hyperparameters are based on experience or marginal likelihood maximization method, which are inaccurate or computationally expensive. In this paper, two posterior methods are proposed to select hyperparameters based on different prior knowledge (constraints), which can obtain the optimal hyperparameters using the optimization theory. Moreover, we also give the theoretical optimal constraints, and verify its effectiveness. Numerical simulation shows that the hyperparameters and parameter vector estimate obtained by the proposed methods are the optimal ones.
文摘Analyzing big data, especially medical data, helps to provide good health care to patients and face the risks of death. The COVID-19 pandemic has had a significant impact on public health worldwide, emphasizing the need for effective risk prediction models. Machine learning (ML) techniques have shown promise in analyzing complex data patterns and predicting disease outcomes. The accuracy of these techniques is greatly affected by changing their parameters. Hyperparameter optimization plays a crucial role in improving model performance. In this work, the Particle Swarm Optimization (PSO) algorithm was used to effectively search the hyperparameter space and improve the predictive power of the machine learning models by identifying the optimal hyperparameters that can provide the highest accuracy. A dataset with a variety of clinical and epidemiological characteristics linked to COVID-19 cases was used in this study. Various machine learning models, including Random Forests, Decision Trees, Support Vector Machines, and Neural Networks, were utilized to capture the complex relationships present in the data. To evaluate the predictive performance of the models, the accuracy metric was employed. The experimental findings showed that the suggested method of estimating COVID-19 risk is effective. When compared to baseline models, the optimized machine learning models performed better and produced better results.
文摘Background:The existence of doublets in single-cell RNA sequencing(scRNA-seq)data poses a great challenge in downstream data analysis.Computational doublet-detection methods have been developed to remove doublets from scRNA-seq data.Yet,the default hyperparameter settings of those methods may not provide optimal performance.Methods:We propose a strategy to tune hyperparameters for a cutting-edge doublet-detection method.We utilize a full factorial design to explore the relationship between hyperparameters and detection accuracy on 16 real scRNA-seq datasets.The optimal hyperparameters are obtained by a response surface model and convex optimization.Results:We show that the optimal hyperparameters provide top performance across scRNA-seq datasets under various biological conditions.Our tuning strategy can be applied to other computational doublet-detection methods.It also offers insights into hyperparameter tuning for broader computational methods in scRNA-seq data analysis.Conclusions:The hyperparameter configuration significantly impacts the performance of computational doublet-detection methods.Our study is the first attempt to systematically explore the optimal hyperparameters under various biological conditions and optimization objectives.Our study provides much-needed guidance for hyperparameter tuning in computational doublet-detection methods.
基金funded by the National Natural Science Foundation of China(No.52376039 and U24A20138)the Beijing Natural Science Foundation of China(No.JQ24017)+1 种基金the National Science and Technology Major Project of China(Nos.J2019-II-0005-0025 and Y2022-Ⅱ-0002-0005)the Special Fund for the Member of Youth Innovation Promotion Association of Chinese Academy of Sciences(No.2018173)。
文摘To predict stall and surge in advance that make the aero-engine compressor operatesafely,a stall prediction model based on deep learning theory is established in the current study.The Long Short-Term Memory(LSTM)originating from the recurrent neural network is used,and a set of measured dynamic pressure datasets including the stall process is used to learn whatdetermines the weight of neural network nodes.Subsequently,the structure and function hyperpa-rameters in the model are deeply optimized,and a set of measured pressure data is used to verify theprediction effects of the model.On this basis of the above good predictive capability,stall in low-and high-speed compressor are predicted by using the established model.When a period of non-stallpressure data is used as input in the model,the model can quickly complete the prediction of sub-sequent time series data through the self-learning and prediction mechanism.Comparison with thereal-time measured pressure data demonstrates that the starting point of the predicted stall is basi-cally the same as that of the measured stall,and the stall can be predicted more than 1 s in advanceso that the occurrence of stall can be avoided.The model of stall prediction in the current study canmake up for the uncertainty of threshold selection of the existing stall warning methods based onmeasured data signal processing.It has a great application potential to predict the stall occurrenceof aero-engine compressor in advance and avoid the accidents.
基金funded by the National Key Research and Development Program of China,grant number:2023YFF0615404.
文摘To overcome the challenges associated with predicting gas extraction performance and mitigating the gradual decline in extraction volume,which adversely impacts gas utilization efficiency in mines,a gas extraction pure volume prediction model was developed using Support Vector Regression(SVR)and Random Forest(RF),with hyperparameters fine-tuned via the Genetic Algorithm(GA).Building upon this,an adaptive control model for gas extraction negative pressure was formulated to maximize the extracted gas volume within the pipeline network,followed by field validation experiments.Experimental results indicate that the GA-SVR model surpasses comparable models in terms of mean absolute error,root mean square error,and mean absolute percentage error.In the extraction process of bedding boreholes,the influence of negative pressure on gas extraction concentration diminishes over time,yet it remains a critical factor in determining the extracted pure volume.In contrast,throughout the entire extraction period of cross-layer boreholes,both extracted pure volume and concentration exhibit pronounced sensitivity to fluctuations in extraction negative pressure.Field experiments demonstrated that the adaptive controlmodel enhanced the average extracted gas volume by 5.08% in the experimental borehole group compared to the control group during the later extraction stage,with a more pronounced increase of 7.15% in the first 15 days.The research findings offer essential technical support for the efficient utilization and long-term sustainable development of mine gas resources.The research findings offer essential technical support for gas disaster mitigation and the sustained,efficient utilization of mine gas.
文摘Traffic forecasting with high precision aids Intelligent Transport Systems(ITS)in formulating and optimizing traffic management strategies.The algorithms used for tuning the hyperparameters of the deep learning models often have accurate results at the expense of high computational complexity.To address this problem,this paper uses the Tree-structured Parzen Estimator(TPE)to tune the hyperparameters of the Long Short-term Memory(LSTM)deep learning framework.The Tree-structured Parzen Estimator(TPE)uses a probabilistic approach with an adaptive searching mechanism by classifying the objective function values into good and bad samples.This ensures fast convergence in tuning the hyperparameter values in the deep learning model for performing prediction while still maintaining a certain degree of accuracy.It also overcomes the problem of converging to local optima and avoids timeconsuming random search and,therefore,avoids high computational complexity in prediction accuracy.The proposed scheme first performs data smoothing and normalization on the input data,which is then fed to the input of the TPE for tuning the hyperparameters.The traffic data is then input to the LSTM model with tuned parameters to perform the traffic prediction.The three optimizers:Adaptive Moment Estimation(Adam),Root Mean Square Propagation(RMSProp),and Stochastic Gradient Descend with Momentum(SGDM)are also evaluated for accuracy prediction and the best optimizer is then chosen for final traffic prediction in TPE-LSTM model.Simulation results verify the effectiveness of the proposed model in terms of accuracy of prediction over the benchmark schemes.
文摘In radiology,magnetic resonance imaging(MRI)is an essential diagnostic tool that provides detailed images of a patient’s anatomical and physiological structures.MRI is particularly effective for detecting soft tissue anomalies.Traditionally,radiologists manually interpret these images,which can be labor-intensive and time-consuming due to the vast amount of data.To address this challenge,machine learning,and deep learning approaches can be utilized to improve the accuracy and efficiency of anomaly detection in MRI scans.This manuscript presents the use of the Deep AlexNet50 model for MRI classification with discriminative learning methods.There are three stages for learning;in the first stage,the whole dataset is used to learn the features.In the second stage,some layers of AlexNet50 are frozen with an augmented dataset,and in the third stage,AlexNet50 with an augmented dataset with the augmented dataset.This method used three publicly available MRI classification datasets:Harvard whole brain atlas(HWBA-dataset),the School of Biomedical Engineering of Southern Medical University(SMU-dataset),and The National Institute of Neuroscience and Hospitals brain MRI dataset(NINS-dataset)for analysis.Various hyperparameter optimizers like Adam,stochastic gradient descent(SGD),Root mean square propagation(RMS prop),Adamax,and AdamW have been used to compare the performance of the learning process.HWBA-dataset registers maximum classification performance.We evaluated the performance of the proposed classification model using several quantitative metrics,achieving an average accuracy of 98%.
基金supported by the MSIT(Ministry of Science and ICT),Republic of Korea,under the ITRC(Information Technology Research Center)Support Program(IITP-2024-RS-2022-00156354)supervised by the IITP(Institute for Information&Communications Technology Planning&Evaluation)supported by the Technology Development Program(RS-2023-00264489)funded by the Ministry of SMEs and Startups(MSS,Republic of Korea).
文摘Fire can cause significant damage to the environment,economy,and human lives.If fire can be detected early,the damage can be minimized.Advances in technology,particularly in computer vision powered by deep learning,have enabled automated fire detection in images and videos.Several deep learning models have been developed for object detection,including applications in fire and smoke detection.This study focuses on optimizing the training hyperparameters of YOLOv8 andYOLOv10models usingBayesianTuning(BT).Experimental results on the large-scale D-Fire dataset demonstrate that this approach enhances detection performance.Specifically,the proposed approach improves the mean average precision at an Intersection over Union(IoU)threshold of 0.5(mAP50)of the YOLOv8s,YOLOv10s,YOLOv8l,and YOLOv10lmodels by 0.26,0.21,0.84,and 0.63,respectively,compared tomodels trainedwith the default hyperparameters.The performance gains are more pronounced in larger models,YOLOv8l and YOLOv10l,than in their smaller counterparts,YOLOv8s and YOLOv10s.Furthermore,YOLOv8 models consistently outperform YOLOv10,with mAP50 improvements of 0.26 for YOLOv8s over YOLOv10s and 0.65 for YOLOv8l over YOLOv10l when trained with BT.These results establish YOLOv8 as the preferred model for fire detection applications where detection performance is prioritized.
基金funded by Anhui NARI ZT Electric Co.,Ltd.,entitled“Research on the Shared Operation and Maintenance Service Model for Metering Equipment and Platform Development for the Modern Industrial Chain”(Grant No.524636250005).
文摘With the rapid adoption of artificial intelligence(AI)in domains such as power,transportation,and finance,the number of machine learning and deep learning models has grown exponentially.However,challenges such as delayed retraining,inconsistent version management,insufficient drift monitoring,and limited data security still hinder efficient and reliable model operations.To address these issues,this paper proposes the Intelligent Model Lifecycle Management Algorithm(IMLMA).The algorithm employs a dual-trigger mechanism based on both data volume thresholds and time intervals to automate retraining,and applies Bayesian optimization for adaptive hyperparameter tuning to improve performance.A multi-metric replacement strategy,incorporating MSE,MAE,and R2,ensures that new models replace existing ones only when performance improvements are guaranteed.A versioning and traceability database supports comparison and visualization,while real-time monitoring with stability analysis enables early warnings of latency and drift.Finally,hash-based integrity checks secure both model files and datasets.Experimental validation in a power metering operation scenario demonstrates that IMLMA reduces model update delays,enhances predictive accuracy and stability,and maintains low latency under high concurrency.This work provides a practical,reusable,and scalable solution for intelligent model lifecycle management,with broad applicability to complex systems such as smart grids.
基金supported by the National Natural Science Foundation of China(51774199)the project of the educational department of Liaoning Province(No LJKMZ20220825).
文摘The 91 measured values of the development height of the water-conducting fracture zone(WCFZ)in deep and thick coal seam mining faces under thick loose layer conditions were collected.Five key characteristic variables influencing the WCFZ height were identified.After removing outliers from the dataset,a Random Forest(RF)regression model optimized by the Sparrow Search Algorithm(SSA)was constructed.The hyperparameters of the RF model were iteratively optimized by minimizing the Out-of-Bag(OOB)error,resulting in the rapid deter-mination of optimal parameters.Specifically,the SSA-RF model achieved an OOB error of 0.148,with 20 de-cision trees,a maximum depth of 8,a minimum split sample size of 2,and a minimum leaf node sample size of 1.Cross-validation experiments were performed using the trained optimal model and compared against other prediction methods.The results showed that the mining height had the most significant correlation with the development height of the WCFZ.The SSA-RF model outperformed all other models,with R2 values exceeding 0.9 across the training,validation,and test datasets.Compared to other models,the SSA-RF model demonstrates a simpler structure,stronger fitting capacity,higher predictive accuracy,and superior stability and generaliza-tion ability.It also exhibits the smallest variation in relative error across datasets,indicating excellent adapt-ability to different data conditions.Furthermore,a numerical model was developed using the hydrogeological data from the 1305 working face at Wanfukou Coal Mine,Shandong Province,China,to simulate the dynamic development of the WCFZ during mining.The SSA-RF model predicted the WCFZ height to be 69.7 m,closely aligning with the PFC2D simulation result of 65 m,with an error of less than 5%.Compared to traditional methods and numerical simulations,the SSA-RF model provides more accurate predictions,showing only a 7.23% deviation from the PFC2D simulation,while traditional empirical formulas yield deviations as large as 19.97%.These results demonstrate the SSA-RF model’s superior predictive capability,reinforcing its reliability and engineering applicability for real-world mining operations.This model holds significant potential for enhancing mining safety and optimizing planning processes,offering a more accurate and efficient approach for WCFZ height prediction.
基金Project supported by the National Natural Science Foundation of China(Nos.12172291,12472357,and 12232015)the Shaanxi Province Outstanding Youth Fund Project(No.2024JC-JCQN-05)the 111 Project(No.BP0719007)。
文摘Sparse identification of nonlinear dynamics(SINDy)has made significant progress in data-driven dynamics modeling.However,determining appropriate hyperparameters and addressing the time-consuming symbolic regression process remain substantial challenges.This study proposes the adaptive backward stepwise selection of fast SINDy(ABSS-FSINDy),which integrates statistical learning-based estimation and technical advancements to significantly reduce simulation time.This approach not only provides insights into the conditions under which SINDy performs optimally but also highlights potential failure points,particularly in the context of backward stepwise selection(BSS).By decoding predefined features into textual expressions,ABSS-FSINDy significantly reduces the simulation time compared with conventional symbolic regression methods.We validate the proposed method through a series of numerical experiments involving both planar/spatial dynamics and high-dimensional chaotic systems,including Lotka-Volterra,hyperchaotic Rossler,coupled Lorenz,and Lorenz 96 benchmark systems.The experimental results demonstrate that ABSS-FSINDy autonomously determines optimal hyperparameters within the SINDy framework,overcoming the curse of dimensionality in high-dimensional simulations.This improvement is substantial across both lowand high-dimensional systems,yielding efficiency gains of one to three orders of magnitude.For instance,in a 20D dynamical system,the simulation time is reduced from 107.63 s to just 0.093 s,resulting in a 3-order-of-magnitude improvement in simulation efficiency.This advancement broadens the applicability of SINDy for the identification and reconstruction of high-dimensional dynamical systems.
基金This research was funded by the Deanship of Scientific Research at Princess Nourah Bint Abdulrahman University through the Fast-track Research Funding Program.
文摘Recently,deep learning(DL)became one of the essential tools in bioinformatics.A modified convolutional neural network(CNN)is employed in this paper for building an integratedmodel for deoxyribonucleic acid(DNA)classification.In any CNN model,convolutional layers are used to extract features followed by max-pooling layers to reduce the dimensionality of features.A novel method based on downsampling and CNNs is introduced for feature reduction.The downsampling is an improved form of the existing pooling layer to obtain better classification accuracy.The two-dimensional discrete transform(2D DT)and two-dimensional random projection(2D RP)methods are applied for downsampling.They convert the high-dimensional data to low-dimensional data and transform the data to the most significant feature vectors.However,there are parameters which directly affect how a CNN model is trained.In this paper,some issues concerned with the training of CNNs have been handled.The CNNs are examined by changing some hyperparameters such as the learning rate,size of minibatch,and the number of epochs.Training and assessment of the performance of CNNs are carried out on 16S rRNA bacterial sequences.Simulation results indicate that the utilization of a CNN based on wavelet subsampling yields the best trade-off between processing time and accuracy with a learning rate equal to 0.0001,a size of minibatch equal to 64,and a number of epochs equal to 20.
基金partially supported by the Medical Research Council Confidence in Concept Award,UK(MC_PC_17171)Royal Society International Exchanges Cost Share Award,UK(RP202G0230)+6 种基金British Heart Foundation Accelerator Award,UK(AA\18\3\34220)Hope Foundation for Cancer Research,UK(RM60G0680)Global Challenges Research Fund(GCRF),UK(P202PF11)Sino-UK Industrial Fund,UK(RP202G0289)LIAS Pioneering Partnerships Award,UK(P202ED10)Data Science Enhancement Fund,UK(P202RE237)Guangxi Key Laboratory of Trusted Software,CN(kx201901).
文摘Since 2019,the coronavirus disease-19(COVID-19)has been spreading rapidly worldwide,posing an unignorable threat to the global economy and human health.It is a disease caused by severe acute respiratory syndrome coronavirus 2,a single-stranded RNA virus of the genus Betacoronavirus.This virus is highly infectious and relies on its angiotensin-converting enzyme 2-receptor to enter cells.With the increase in the number of confirmed COVID-19 diagnoses,the difficulty of diagnosis due to the lack of global healthcare resources becomes increasingly apparent.Deep learning-based computer-aided diagnosis models with high generalisability can effectively alleviate this pressure.Hyperparameter tuning is essential in training such models and significantly impacts their final performance and training speed.However,traditional hyperparameter tuning methods are usually time-consuming and unstable.To solve this issue,we introduce Particle Swarm Optimisation to build a PSO-guided Self-Tuning Convolution Neural Network(PSTCNN),allowing the model to tune hyperparameters automatically.Therefore,the proposed approach can reduce human involvement.Also,the optimisation algorithm can select the combination of hyperparameters in a targeted manner,thus stably achieving a solution closer to the global optimum.Experimentally,the PSTCNN can obtain quite excellent results,with a sensitivity of 93.65%±1.86%,a specificity of 94.32%±2.07%,a precision of 94.30%±2.04%,an accuracy of 93.99%±1.78%,an F1-score of 93.97%±1.78%,Matthews Correlation Coefficient of 87.99%±3.56%,and Fowlkes-Mallows Index of 93.97%±1.78%.Our experiments demonstrate that compared to traditional methods,hyperparameter tuning of the model using an optimisation algorithm is faster and more effective.
基金This research was supported by the Universiti Malaya Impact-oriented Interdisciplinary Research Grant Programme(IIRG)-IIRG002C-19HWBUniversiti Malaya Covid-19 Related Special Research Grant(UMCSRG)CSRG008-2020ST and Partnership Grant(RK012-2019)from University of Malaya.
文摘Abnormalities of the gastrointestinal tract are widespread worldwide today.Generally,an effective way to diagnose these life-threatening diseases is based on endoscopy,which comprises a vast number of images.However,the main challenge in this area is that the process is time-consuming and fatiguing for a gastroenterologist to examine every image in the set.Thus,this led to the rise of studies on designingAI-based systems to assist physicians in the diagnosis.In several medical imaging tasks,deep learning methods,especially convolutional neural networks(CNNs),have contributed to the stateof-the-art outcomes,where the complicated nonlinear relation between target classes and data can be learned and not limit to hand-crafted features.On the other hand,hyperparameters are commonly set manually,which may take a long time and leave the risk of non-optimal hyperparameters for classification.An effective tool for tuning optimal hyperparameters of deep CNNis Bayesian optimization.However,due to the complexity of the CNN,the network can be regarded as a black-box model where the information stored within it is hard to interpret.Hence,Explainable Artificial Intelligence(XAI)techniques are applied to overcome this issue by interpreting the decisions of the CNNs in such wise the physicians can trust.To play an essential role in real-time medical diagnosis,CNN-based models need to be accurate and interpretable,while the uncertainty must be handled.Therefore,a novel method comprising of three phases is proposed to classify these life-threatening diseases.At first,hyperparameter tuning is performed using Bayesian optimization for two state-of-the-art deep CNNs,and then Darknet53 and InceptionV3 features are extracted from these fine-tunned models.Secondly,XAI techniques are used to interpret which part of the images CNN takes for feature extraction.At last,the features are fused,and uncertainties are handled by selecting entropybased features.The experimental results show that the proposed method outperforms existing methods by achieving an accuracy of 97%based on a Bayesian optimized Support Vector Machine classifier.
文摘With the advancement of artificial intelligence,traffic forecasting is gaining more and more interest in optimizing route planning and enhancing service quality.Traffic volume is an influential parameter for planning and operating traffic structures.This study proposed an improved ensemble-based deep learning method to solve traffic volume prediction problems.A set of optimal hyperparameters is also applied for the suggested approach to improve the performance of the learning process.The fusion of these methodologies aims to harness ensemble empirical mode decomposition’s capacity to discern complex traffic patterns and long short-term memory’s proficiency in learning temporal relationships.Firstly,a dataset for automatic vehicle identification is obtained and utilized in the preprocessing stage of the ensemble empirical mode decomposition model.The second aspect involves predicting traffic volume using the long short-term memory algorithm.Next,the study employs a trial-and-error approach to select a set of optimal hyperparameters,including the lookback window,the number of neurons in the hidden layers,and the gradient descent optimization.Finally,the fusion of the obtained results leads to a final traffic volume prediction.The experimental results show that the proposed method outperforms other benchmarks regarding various evaluation measures,including mean absolute error,root mean squared error,mean absolute percentage error,and R-squared.The achieved R-squared value reaches an impressive 98%,while the other evaluation indices surpass the competing.These findings highlight the accuracy of traffic pattern prediction.Consequently,this offers promising prospects for enhancing transportation management systems and urban infrastructure planning.
基金The authors would like to acknowledge the support of the Deputy for Research and Innovation,Ministry of Education,Kingdom of Saudi Arabia for this research through a grant(NU/IFC/INT/01/008)under the institutional Funding Committee at Najran University,Kingdom of Saudi Arabia.
文摘Deep learning techniques,particularly convolutional neural networks(CNNs),have exhibited remarkable performance in solving visionrelated problems,especially in unpredictable,dynamic,and challenging environments.In autonomous vehicles,imitation-learning-based steering angle prediction is viable due to the visual imagery comprehension of CNNs.In this regard,globally,researchers are currently focusing on the architectural design and optimization of the hyperparameters of CNNs to achieve the best results.Literature has proven the superiority of metaheuristic algorithms over the manual-tuning of CNNs.However,to the best of our knowledge,these techniques are yet to be applied to address the problem of imitationlearning-based steering angle prediction.Thus,in this study,we examine the application of the bat algorithm and particle swarm optimization algorithm for the optimization of the CNN model and its hyperparameters,which are employed to solve the steering angle prediction problem.To validate the performance of each hyperparameters’set and architectural parameters’set,we utilized the Udacity steering angle dataset and obtained the best results at the following hyperparameter set:optimizer,Adagrad;learning rate,0.0052;and nonlinear activation function,exponential linear unit.As per our findings,we determined that the deep learning models show better results but require more training epochs and time as compared to shallower ones.Results show the superiority of our approach in optimizing CNNs through metaheuristic algorithms as compared with the manual-tuning approach.Infield testing was also performed using the model trained with the optimal architecture,which we developed using our approach.
文摘Hydrological models are developed to simulate river flows over a watershed for many practical applications in the field of water resource management. The present paper compares the performance of two recurrent neural networks for rainfall-runoff modeling in the Zou River basin at Atchérigbé outlet. To this end, we used daily precipitation data over the period 1988-2010 as input of the models, such as the Long Short-Term Memory (LSTM) and Recurrent Gate Networks (GRU) to simulate river discharge in the study area. The investigated models give good results in calibration (R2 = 0.888, NSE = 0.886, and RMSE = 0.42 for LSTM;R2 = 0.9, NSE = 0.9 and RMSE = 0.397 for GRU) and in validation (R2 = 0.865, NSE = 0.851, and RMSE = 0.329 for LSTM;R2 = 0.9, NSE = 0.865 and RMSE = 0.301 for GRU). This good performance of LSTM and GRU models confirms the importance of models based on machine learning in modeling hydrological phenomena for better decision-making.
文摘Potato late blight and early blight are common hazards to the long-term production of potatoes, impacting many farmers around the world, particularly in Africa. Early detection and treatment of the potato blight disease are critical for promoting healthy potato plant growth and ensuring adequate supply and food security for the fast-growing population. As a result, machine-driven disease detection systems may be able to overcome the constraints of traditional leaf disease diagnosis procedures, which are generally time-consuming, inaccurate, and costly. Convolutional Neural Networks (CNNs) have been shown to be effective in a variety of agricultural applications. CNNs have been shown to be helpful in detecting disease in plants because of their capacity to analyze vast volumes of data quickly and reliably. However, the method hasn’t been widely used in the detection of potato late blight and early blight diseases, which reduce yields significantly. The goal of this study was to compare six cutting-edge CNN architectural models, taking into account transfer learning for training and four hyperparameters. The CNN architectures evaluated were AlexNet, GoogleNet, SqueezeNet, DenseNet121, EfficientNet b7, and VGG19. Likewise, the hyperparameters analyzed were the number of epochs, the batch size, the optimizer, and the learning rate. An open-source dataset containing 4082 images was used. The DenseNet121 architecture with a batch of 32 and a Stochastic Gradient Descent (SGD) optimizer with a learning rate of 0.01 produced the best performance, with an accuracy of 98.34% and a 97.37% f1-score. The DenseNet121 model was shown to be useful in developing computer vision systems that aid farmers in improving their disease management systems for potato cultivation.
基金supported in part by the National Natural Science Foundation of China under Grant No.61503059
文摘Hyperparameters are important for machine learning algorithms since they directly control the behaviors of training algorithms and have a significant effect on the performance of machine learning models. Several techniques have been developed and successfully applied for certain application domains. However, this work demands professional knowledge and expert experience. And sometimes it has to resort to the brute-force search.Therefore, if an efficient hyperparameter optimization algorithm can be developed to optimize any given machine learning method, it will greatly improve the efficiency of machine learning. In this paper, we consider building the relationship between the performance of the machine learning models and their hyperparameters by Gaussian processes. In this way, the hyperparameter tuning problem can be abstracted as an optimization problem and Bayesian optimization is used to solve the problem. Bayesian optimization is based on the Bayesian theorem. It sets a prior over the optimization function and gathers the information from the previous sample to update the posterior of the optimization function. A utility function selects the next sample point to maximize the optimization function.Several experiments were conducted on standard test datasets. Experiment results show that the proposed method can find the best hyperparameters for the widely used machine learning models, such as the random forest algorithm and the neural networks, even multi-grained cascade forest under the consideration of time cost.
