Software defect prediction aims to use measurement data of code and historical defects to predict potential problems,optimize testing resources and defect management.However,current methods face challenges:(1)Coarse-g...Software defect prediction aims to use measurement data of code and historical defects to predict potential problems,optimize testing resources and defect management.However,current methods face challenges:(1)Coarse-grained file level detection cannot accurately locate specific defects.(2)Fine-grained line-level defect prediction methods rely solely on local information of a single line of code,failing to deeply analyze the semantic context of the code line and ignoring the heuristic impact of line-level context on the code line,making it difficult to capture the interaction between global and local information.Therefore,this paper proposes a telecontext-enhanced recursive interactive attention fusion method for line-level defect prediction(TRIA-LineDP).Firstly,using a bidirectional hierarchical attention network to extract semantic features and contextual information from the original code lines as the basis.Then,the extracted contextual information is forwarded to the telecontext capture module to aggregate the global context,thereby enhancing the understanding of broader code dynamics.Finally,a recursive interaction model is used to simulate the interaction between code lines and line-level context,passing information layer by layer to enhance local and global information exchange,thereby achieving accurate defect localization.Experimental results from within-project defect prediction(WPDP)and cross-project defect prediction(CPDP)conducted on nine different projects(encompassing a total of 32 versions)demonstrated that,within the same project,the proposed methods will respectively recall at top 20%of lines of code(Recall@Top20%LOC)and effort at top 20%recall(Effort@Top20%Recall)has increased by 11%–52%and 23%–77%.In different projects,improvements of 9%–60%and 18%–77%have been achieved,which are superior to existing advanced methods and have good detection performance.展开更多
Accurately predicting environmental parameters in solar greenhouses is crucial for achieving precise environmental control.In solar greenhouses,temperature,humidity,and light intensity are crucial environmental parame...Accurately predicting environmental parameters in solar greenhouses is crucial for achieving precise environmental control.In solar greenhouses,temperature,humidity,and light intensity are crucial environmental parameters.The monitoring platform collected data on the internal environment of the solar greenhouse for one year,including temperature,humidity,and light intensity.Additionally,meteorological data,comprising outdoor temperature,outdoor humidity,and outdoor light intensity,was gathered during the same time frame.The characteristics and interrelationships among these parameters were investigated by a thorough analysis.The analysis revealed that environmental parameters in solar greenhouses displayed characteristics such as temporal variability,non-linearity,and periodicity.These parameters exhibited complex coupling relationships.Notably,these characteristics and coupling relationships exhibited pronounced seasonal variations.The multi-parameter multi-step prediction model for solar greenhouse(MPMS-SGH)was introduced,aiming to accurately predict three key greenhouse environmental parameters,and the model had certain seasonal adaptability.MPMS-SGH was structured with multiple layers,including an input layer,a preprocessing layer,a feature extraction layer,and a prediction layer.The input layer was used to generate the original sequence matrix,which included indoor temperature,indoor humidity,indoor light intensity,as well as outdoor temperature and outdoor light intensity.Then the preprocessing layer normalized,decomposed,and positionally encoded the original sequence matrix.In the feature extraction layer,the time attention mechanism and frequency attention mechanism were used to extract features from the trend component and the seasonal component,respectively.Finally,the prediction layer used a multi-layer perceptron to perform multi-step prediction of indoor environmental parameters(i.e.temperature,humidity,and light intensity).The parameter selection experiment evaluated the predictive performance of MPMS-SGH on input and output sequences of different lengths.The results indicated that with a constant output sequence length,the prediction accuracy of MPMS-SGH was firstly increased and then decreased with the increase of input sequence length.Specifically,when the input sequence length was 100,MPMS-SGH had the highest prediction accuracy,with RMSE of 0.22℃,0.28%,and 250lx for temperature,humidity,and light intensity,respectively.When the length of the input sequence remained constant,as the length of the output sequence increased,the accuracy of the model in predicting the three environmental parameters was continuously decreased.When the length of the output sequence exceeded 45,the prediction accuracy of MPMS-SGH was significantly decreased.In order to achieve the best balance between model size and performance,the input sequence length of MPMS-SGH was set to be 100,while the output sequence length was set to be 35.To assess MPMS-SGH’s performance,comparative experiments with four prediction models were conducted:SVR,STL-SVR,LSTM,and STL-LSTM.The results demonstrated that MPMS-SGH surpassed all other models,achieving RMSE of 0.15℃for temperature,0.38%for humidity,and 260lx for light intensity.Additionally,sequence decomposition can contribute to enhancing MPMS-SGH’s prediction performance.To further evaluate MPMS-SGH’s capabilities,its prediction accuracy was tested across different seasons for greenhouse environmental parameters.MPMS-SGH had the highest accuracy in predicting indoor temperature and the lowest accuracy in predicting humidity.And the accuracy of MPMS-SGH in predicting environmental parameters of the solar greenhouse fluctuated with seasons.MPMS-SGH had the highest accuracy in predicting the temperature inside the greenhouse on sunny days in spring(R^(2)=0.91),the highest accuracy in predicting the humidity inside the greenhouse on sunny days in winter(R^(2)=0.83),and the highest accuracy in predicting the light intensity inside the greenhouse on cloudy days in autumm(R^(2)=0.89).MPMS-SGH had the lowest accuracy in predicting three environmental parameters in a sunny summer greenhouse.展开更多
Accurate short-term photovoltaic(PV)output forecasting is beneficial for increasing grid stabil-ity and enhancing the capacity for photovoltaic power absorption.In response to the challenges faced by commonly used pho...Accurate short-term photovoltaic(PV)output forecasting is beneficial for increasing grid stabil-ity and enhancing the capacity for photovoltaic power absorption.In response to the challenges faced by commonly used photovoltaic forecasting methods,which struggle to handle issues such as non-u-niform lengths of time series data for power generation and meteorological conditions,overlapping photovoltaic characteristics,and nonlinear correlations,an improved method that utilizes spectral clustering and dynamic time warping(DTW)for selecting similar days is proposed to optimize the dataset along the temporal dimension.Furthermore,XGBoost is employed for recursive feature selec-tion.On this basis,to address the issue that single forecasting models excel at capturing different data characteristics and tend to exhibit significant prediction errors under adverse meteorological con-ditions,an improved forecasting model based on Stacking and weighted fusion is proposed to reduce the independent bias and variance of individual models and enhance the predictive accuracy.Final-ly,experimental validation is carried out using real data from a photovoltaic power station in the Xi-aoshan District of Hangzhou,China,demonstrating that the proposed method can still achieve accu-rate and robust forecasting results even under conditions of significant meteorological fluctuations.展开更多
In this study,the hourly directions of eight banking stocks in Borsa Istanbul were predicted using linear-based,deep-learning(LSTM)and ensemble learning(Light-GBM)models.These models were trained with four different f...In this study,the hourly directions of eight banking stocks in Borsa Istanbul were predicted using linear-based,deep-learning(LSTM)and ensemble learning(Light-GBM)models.These models were trained with four different feature sets and their performances were evaluated in terms of accuracy and F-measure metrics.While the first experiments directly used the own stock features as the model inputs,the second experiments utilized reduced stock features through Variational AutoEncoders(VAE).In the last experiments,in order to grasp the effects of the other banking stocks on individual stock performance,the features belonging to other stocks were also given as inputs to our models.While combining other stock features was done for both own(named as allstock_own)and VAE-reduced(named as allstock_VAE)stock features,the expanded dimensions of the feature sets were reduced by Recursive Feature Elimination.As the highest success rate increased up to 0.685 with allstock_own and LSTM with attention model,the combination of allstock_VAE and LSTM with the attention model obtained an accuracy rate of 0.675.Although the classification results achieved with both feature types was close,allstock_VAE achieved these results using nearly 16.67%less features compared to allstock_own.When all experimental results were examined,it was found out that the models trained with allstock_own and allstock_VAE achieved higher accuracy rates than those using individual stock features.It was also concluded that the results obtained with the VAE-reduced stock features were similar to those obtained by own stock features.展开更多
This paper investigates the characteristics of a non-stationary time series, which exists in mechanical fault diagnosis. Combining the characteristics with predictive efficiency, the limitation of the ARIMA model pred...This paper investigates the characteristics of a non-stationary time series, which exists in mechanical fault diagnosis. Combining the characteristics with predictive efficiency, the limitation of the ARIMA model prediction method is analyzed. This model often is applied in the prediction of a non-stationary times series in present. Thus, a wavelet prediction method is introduced to solve non-stationary problems. The Mallat method, often used in signal processing, results form the decimation or the retention of one out of every two samples. Its advantage is that just enough information is kept to allow the exact reconstruction of the input series, but the disadvantage is a time-varying series on line cannot be pursued. Therefore, the authors present another method, à Trous method, which can be applied for recursive prediction in real-time sampling procedure.展开更多
In the strip rolling process, shape control system possesses the characteristics of nonlinearity, strong coupling, time delay and time variation. Based on self adapting Elman dynamic recursion network prediction model...In the strip rolling process, shape control system possesses the characteristics of nonlinearity, strong coupling, time delay and time variation. Based on self adapting Elman dynamic recursion network prediction model, the fuzzy control method was used to control the shape on four-high cold mill. The simulation results showed that the system can be applied to real time on line control of the shape.展开更多
Traffic flow prediction is an important part of the intelligent transportation system. Accurate multi-step traffic flow prediction plays an important role in improving the operational efficiency of the traffic network...Traffic flow prediction is an important part of the intelligent transportation system. Accurate multi-step traffic flow prediction plays an important role in improving the operational efficiency of the traffic network. Since traffic flow data has complex spatio-temporal correlation and non-linearity, existing prediction methods are mainly accomplished through a combination of a Graph Convolutional Network (GCN) and a recurrent neural network. The combination strategy has an excellent performance in traffic prediction tasks. However, multi-step prediction error accumulates with the predicted step size. Some scholars use multiple sampling sequences to achieve more accurate prediction results. But it requires high hardware conditions and multiplied training time. Considering the spatiotemporal correlation of traffic flow and influence of external factors, we propose an Attention Based Spatio-Temporal Graph Convolutional Network considering External Factors (ABSTGCN-EF) for multi-step traffic flow prediction. This model models the traffic flow as diffusion on a digraph and extracts the spatial characteristics of traffic flow through GCN. We add meaningful time-slots attention to the encoder-decoder to form an Attention Encoder Network (AEN) to handle temporal correlation. The attention vector is used as a competitive choice to draw the correlation between predicted states and historical states. We considered the impact of three external factors (daytime, weekdays, and traffic accident markers) on the traffic flow prediction tasks. Experiments on two public data sets show that it makes sense to consider external factors. The prediction performance of our ABSTGCN-EF model achieves 7.2%–8.7% higher than the state-of-the-art baselines.展开更多
Considering chaotic time series multi-step prediction, multi-step direct prediction model based on partial least squares （PLS） is proposed in this article, where PLS, the method for predicting a set of dependent var...Considering chaotic time series multi-step prediction, multi-step direct prediction model based on partial least squares （PLS） is proposed in this article, where PLS, the method for predicting a set of dependent variables forming a large set of predictors, is used to model the dynamic evolution between the space points and the corresponding future points. The model can eliminate error accumulation with the common single-step local model algorithm~ and refrain from the high multi-collinearity problem in the reconstructed state space with the increase of embedding dimension. Simulation predictions are done on the Mackey-Glass chaotic time series with the model. The satisfying prediction accuracy is obtained and the model efficiency verified. In the experiments, the number of extracted components in PLS is set with cross-validation procedure.展开更多
A content-aware multi-step prediction control(CAMPC)algorithm is proposed to determine the bitrate of 360-degree videos,aim⁃ing to enhance the quality of experience(QoE)of users and reduce the cost of video content pr...A content-aware multi-step prediction control(CAMPC)algorithm is proposed to determine the bitrate of 360-degree videos,aim⁃ing to enhance the quality of experience(QoE)of users and reduce the cost of video content providers(VCP).The CAMPC algorithm first em⁃ploys a neural network to generate the content richness and combines it with the current field of view(FOV)to accurately predict the probability distribution of tiles being viewed.Then,for the tiles in the predicted viewport which directly affect QoE,the CAMPC algorithm utilizes a multi-step prediction for future system states,and accordingly selects the bitrates of multiple subsequent steps,instead of an instantaneous state.Meanwhile,it controls the buffer occupancy to eliminate the impact of prediction errors.We implement CAMPC on players by building a 360-degree video streaming platform and evaluating other advanced adaptive bitrate(ABR)rules through the real network.Experimental results show that CAMPC can save 83.5%of bandwidth resources compared with the scheme that completely transmits the tiles outside the viewport with the Dynamic Adaptive Streaming over HTTP(DASH)protocol.Besides,the proposed method can improve the system utility by 62.7%and 27.6%compared with the DASH official and viewport-based rules,respectively.展开更多
Virtual machine(VM)consolidation is an effective way to improve resource utilization and reduce energy consumption in cloud data centers.Most existing studies have considered VM consolidation as a bin-packing problem,...Virtual machine(VM)consolidation is an effective way to improve resource utilization and reduce energy consumption in cloud data centers.Most existing studies have considered VM consolidation as a bin-packing problem,but the current schemes commonly ignore the long-term relationship between VMs and hosts.In addition,there is a lack of long-term consideration for resource optimization in the VM consolidation,which results in unnecessary VM migration and increased energy consumption.To address these limitations,a VM consolidation method based on multi-step prediction and affinity-aware technique for energy-efficient cloud data centers(MPaAF-VMC)is proposed.The proposed method uses an improved linear regression prediction algorithm to predict the next-moment resource utilization of hosts and VMs,and obtains the stage demand of resources in the future period through multi-step prediction,which is realized by iterative prediction.Then,based on the multi-step prediction,an affinity model between the VM and host is designed using the first-order correlation coefficient and Euclidean distance.During the VM consolidation,the affinity value is used to select the migration VM and placement host.The proposed method is compared with the existing consolidation algorithms on the PlanetLab and Google cluster real workload data using the CloudSim simulation platform.Experimental results show that the proposed method can achieve significant improvement in reducing energy consumption,VM migration costs,and service level agreement(SLA)violations.展开更多
This paper studied a fast recursive predictive algorithm used for medical X-ray image compression. This algorithm consists of mathematics model building, fast recursive algorithm deducing, initial value determining, s...This paper studied a fast recursive predictive algorithm used for medical X-ray image compression. This algorithm consists of mathematics model building, fast recursive algorithm deducing, initial value determining, step-size selecting, image compression encoding and original image recovering. The experiment result indicates that this algorithm has not only a higher compression ratio to medical X-ray images compression, but also promotes image compression speed greatly.展开更多
The rolling force model for cold tandem mill was put forward by using the Elman dynamic recursive network method,based on the actual measured data.Furthermore,a good assumption is put forward,which brings a full unive...The rolling force model for cold tandem mill was put forward by using the Elman dynamic recursive network method,based on the actual measured data.Furthermore,a good assumption is put forward,which brings a full universe of discourse self-adjusting factor fuzzy control,closed-loop adjusting,based on error feedback and expertise into a rolling force prediction model,to modify prediction outputs and improve prediction precision and robustness.The simulated results indicate that the method is highly effective and the prediction precision is better than that of the traditional method.Predicted relative error is less than ±4%,so the prediction is high precise for the cold tandem mill.展开更多
A compound neural network was constructed during the process of identification and multi-step prediction. Under the PID-type long-range predictive cost function, the control signal was calculated based on gradient alg...A compound neural network was constructed during the process of identification and multi-step prediction. Under the PID-type long-range predictive cost function, the control signal was calculated based on gradient algorithm. The nonlinear controller’s structure was similar to the conventional PID controller. The parameters of this controller were tuned by using a local recurrent neural network on-line. The controller has a better effect than the conventional PID controller. Simulation study shows the effectiveness and good performance.展开更多
For the unforced dynamical non-linear state–space model,a new Q1 and efficient square root extended kernel recursive least square estimation algorithm is developed in this article.The proposed algorithm lends itself ...For the unforced dynamical non-linear state–space model,a new Q1 and efficient square root extended kernel recursive least square estimation algorithm is developed in this article.The proposed algorithm lends itself towards the parallel implementation as in the FPGA systems.With the help of an ortho-normal triangularization method,which relies on numerically stable givens rotation,matrix inversion causes a computational burden,is reduced.Matrix computation possesses many excellent numerical properties such as singularity,symmetry,skew symmetry,and triangularity is achieved by using this algorithm.The proposed method is validated for the prediction of stationary and non-stationary Mackey–Glass Time Series,along with that a component in the x-direction of the Lorenz Times Series is also predicted to illustrate its usefulness.By the learning curves regarding mean square error(MSE)are witnessed for demonstration with prediction performance of the proposed algorithm from where it’s concluded that the proposed algorithm performs better than EKRLS.This new SREKRLS based design positively offers an innovative era towards non-linear systolic arrays,which is efficient in developing very-large-scale integration(VLSI)applications with non-linear input data.Multiple experiments are carried out to validate the reliability,effectiveness,and applicability of the proposed algorithm and with different noise levels compared to the Extended kernel recursive least-squares(EKRLS)algorithm.展开更多
A DRNN (diagonal recurrent neural network) and its RPE (recurrent prediction error) learning algorithm are proposed in this paper .Using of the simple structure of DRNN can reduce the capacity of calculation. The prin...A DRNN (diagonal recurrent neural network) and its RPE (recurrent prediction error) learning algorithm are proposed in this paper .Using of the simple structure of DRNN can reduce the capacity of calculation. The principle of RPE learning algorithm is to adjust weights along the direction of Gauss-Newton. Meanwhile, it is unnecessary to calculate the second local derivative and the inverse matrixes, whose unbiasedness is proved. With application to the extremely short time prediction of large ship pitch, satisfactory results are obtained. Prediction effect of this algorithm is compared with that of auto-regression and periodical diagram method, and comparison results show that the proposed algorithm is feasible.展开更多
Video prediction is the problem of generating future frames by exploiting the spatiotemporal correlation from the past frame sequence.It is one of the crucial issues in computer vision and has many real-world applicat...Video prediction is the problem of generating future frames by exploiting the spatiotemporal correlation from the past frame sequence.It is one of the crucial issues in computer vision and has many real-world applications,mainly focused on predicting future scenarios to avoid undesirable outcomes.However,modeling future image content and object is challenging due to the dynamic evolution and complexity of the scene,such as occlusions,camera movements,delay and illumination.Direct frame synthesis or optical-flow estimation are common approaches used by researchers.However,researchers mainly focused on video prediction using one of the approaches.Both methods have limitations,such as direct frame synthesis,usually face blurry prediction due to complex pixel distributions in the scene,and optical-flow estimation,usually produce artifacts due to large object displacements or obstructions in the clip.In this paper,we constructed a deep neural network Frame Prediction Network(FPNet-OF)with multiplebranch inputs(optical flow and original frame)to predict the future video frame by adaptively fusing the future object-motion with the future frame generator.The key idea is to jointly optimize direct RGB frame synthesis and dense optical flow estimation to generate a superior video prediction network.Using various real-world datasets,we experimentally verify that our proposed framework can produce high-level video frame compared to other state-ofthe-art framework.展开更多
A good machine learning model would greatly contribute to an accurate crime prediction. Thus, researchers select advanced models more frequently than basic models. To find out whether advanced models have a prominent ...A good machine learning model would greatly contribute to an accurate crime prediction. Thus, researchers select advanced models more frequently than basic models. To find out whether advanced models have a prominent advantage, this study focuses shift from obtaining crime prediction to on comparing model performance between these two types of models on crime prediction. In this study, we aimed to predict burglary occurrence in Los Angeles City, and compared a basic model just using prior year burglary occurrence with advanced models including linear regressor and random forest regressor. In addition, American Community Survey data was used to provide neighborhood level socio-economic features. After finishing data preprocessing steps that regularize the dataset, recursive feature elimination was utilized to determine the final features and the parameters of the two advanced models. Finally, to find out the best fit model, three metrics were used to evaluate model performance: R squared, adjusted R squared and mean squared error. The results indicate that linear regressor is the most suitable model among three models applied in the study with a slightly smaller mean squared error than that of basic model, whereas random forest model performed worse than the basic model. With a much more complex learning steps, advanced models did not show prominent advantages, and further research to extend the current study were discussed.展开更多
基金supported by National Natural Science Foundation of China(no.62376240).
文摘Software defect prediction aims to use measurement data of code and historical defects to predict potential problems,optimize testing resources and defect management.However,current methods face challenges:(1)Coarse-grained file level detection cannot accurately locate specific defects.(2)Fine-grained line-level defect prediction methods rely solely on local information of a single line of code,failing to deeply analyze the semantic context of the code line and ignoring the heuristic impact of line-level context on the code line,making it difficult to capture the interaction between global and local information.Therefore,this paper proposes a telecontext-enhanced recursive interactive attention fusion method for line-level defect prediction(TRIA-LineDP).Firstly,using a bidirectional hierarchical attention network to extract semantic features and contextual information from the original code lines as the basis.Then,the extracted contextual information is forwarded to the telecontext capture module to aggregate the global context,thereby enhancing the understanding of broader code dynamics.Finally,a recursive interaction model is used to simulate the interaction between code lines and line-level context,passing information layer by layer to enhance local and global information exchange,thereby achieving accurate defect localization.Experimental results from within-project defect prediction(WPDP)and cross-project defect prediction(CPDP)conducted on nine different projects(encompassing a total of 32 versions)demonstrated that,within the same project,the proposed methods will respectively recall at top 20%of lines of code(Recall@Top20%LOC)and effort at top 20%recall(Effort@Top20%Recall)has increased by 11%–52%and 23%–77%.In different projects,improvements of 9%–60%and 18%–77%have been achieved,which are superior to existing advanced methods and have good detection performance.
文摘Accurately predicting environmental parameters in solar greenhouses is crucial for achieving precise environmental control.In solar greenhouses,temperature,humidity,and light intensity are crucial environmental parameters.The monitoring platform collected data on the internal environment of the solar greenhouse for one year,including temperature,humidity,and light intensity.Additionally,meteorological data,comprising outdoor temperature,outdoor humidity,and outdoor light intensity,was gathered during the same time frame.The characteristics and interrelationships among these parameters were investigated by a thorough analysis.The analysis revealed that environmental parameters in solar greenhouses displayed characteristics such as temporal variability,non-linearity,and periodicity.These parameters exhibited complex coupling relationships.Notably,these characteristics and coupling relationships exhibited pronounced seasonal variations.The multi-parameter multi-step prediction model for solar greenhouse(MPMS-SGH)was introduced,aiming to accurately predict three key greenhouse environmental parameters,and the model had certain seasonal adaptability.MPMS-SGH was structured with multiple layers,including an input layer,a preprocessing layer,a feature extraction layer,and a prediction layer.The input layer was used to generate the original sequence matrix,which included indoor temperature,indoor humidity,indoor light intensity,as well as outdoor temperature and outdoor light intensity.Then the preprocessing layer normalized,decomposed,and positionally encoded the original sequence matrix.In the feature extraction layer,the time attention mechanism and frequency attention mechanism were used to extract features from the trend component and the seasonal component,respectively.Finally,the prediction layer used a multi-layer perceptron to perform multi-step prediction of indoor environmental parameters(i.e.temperature,humidity,and light intensity).The parameter selection experiment evaluated the predictive performance of MPMS-SGH on input and output sequences of different lengths.The results indicated that with a constant output sequence length,the prediction accuracy of MPMS-SGH was firstly increased and then decreased with the increase of input sequence length.Specifically,when the input sequence length was 100,MPMS-SGH had the highest prediction accuracy,with RMSE of 0.22℃,0.28%,and 250lx for temperature,humidity,and light intensity,respectively.When the length of the input sequence remained constant,as the length of the output sequence increased,the accuracy of the model in predicting the three environmental parameters was continuously decreased.When the length of the output sequence exceeded 45,the prediction accuracy of MPMS-SGH was significantly decreased.In order to achieve the best balance between model size and performance,the input sequence length of MPMS-SGH was set to be 100,while the output sequence length was set to be 35.To assess MPMS-SGH’s performance,comparative experiments with four prediction models were conducted:SVR,STL-SVR,LSTM,and STL-LSTM.The results demonstrated that MPMS-SGH surpassed all other models,achieving RMSE of 0.15℃for temperature,0.38%for humidity,and 260lx for light intensity.Additionally,sequence decomposition can contribute to enhancing MPMS-SGH’s prediction performance.To further evaluate MPMS-SGH’s capabilities,its prediction accuracy was tested across different seasons for greenhouse environmental parameters.MPMS-SGH had the highest accuracy in predicting indoor temperature and the lowest accuracy in predicting humidity.And the accuracy of MPMS-SGH in predicting environmental parameters of the solar greenhouse fluctuated with seasons.MPMS-SGH had the highest accuracy in predicting the temperature inside the greenhouse on sunny days in spring(R^(2)=0.91),the highest accuracy in predicting the humidity inside the greenhouse on sunny days in winter(R^(2)=0.83),and the highest accuracy in predicting the light intensity inside the greenhouse on cloudy days in autumm(R^(2)=0.89).MPMS-SGH had the lowest accuracy in predicting three environmental parameters in a sunny summer greenhouse.
基金Supported by the National Natural Science Foundation of China(No.52005442)the Technology Project of Zhejiang Huayun Information Technology Co.,Ltd.(No.HYJT/JS-2020-004).
文摘Accurate short-term photovoltaic(PV)output forecasting is beneficial for increasing grid stabil-ity and enhancing the capacity for photovoltaic power absorption.In response to the challenges faced by commonly used photovoltaic forecasting methods,which struggle to handle issues such as non-u-niform lengths of time series data for power generation and meteorological conditions,overlapping photovoltaic characteristics,and nonlinear correlations,an improved method that utilizes spectral clustering and dynamic time warping(DTW)for selecting similar days is proposed to optimize the dataset along the temporal dimension.Furthermore,XGBoost is employed for recursive feature selec-tion.On this basis,to address the issue that single forecasting models excel at capturing different data characteristics and tend to exhibit significant prediction errors under adverse meteorological con-ditions,an improved forecasting model based on Stacking and weighted fusion is proposed to reduce the independent bias and variance of individual models and enhance the predictive accuracy.Final-ly,experimental validation is carried out using real data from a photovoltaic power station in the Xi-aoshan District of Hangzhou,China,demonstrating that the proposed method can still achieve accu-rate and robust forecasting results even under conditions of significant meteorological fluctuations.
文摘In this study,the hourly directions of eight banking stocks in Borsa Istanbul were predicted using linear-based,deep-learning(LSTM)and ensemble learning(Light-GBM)models.These models were trained with four different feature sets and their performances were evaluated in terms of accuracy and F-measure metrics.While the first experiments directly used the own stock features as the model inputs,the second experiments utilized reduced stock features through Variational AutoEncoders(VAE).In the last experiments,in order to grasp the effects of the other banking stocks on individual stock performance,the features belonging to other stocks were also given as inputs to our models.While combining other stock features was done for both own(named as allstock_own)and VAE-reduced(named as allstock_VAE)stock features,the expanded dimensions of the feature sets were reduced by Recursive Feature Elimination.As the highest success rate increased up to 0.685 with allstock_own and LSTM with attention model,the combination of allstock_VAE and LSTM with the attention model obtained an accuracy rate of 0.675.Although the classification results achieved with both feature types was close,allstock_VAE achieved these results using nearly 16.67%less features compared to allstock_own.When all experimental results were examined,it was found out that the models trained with allstock_own and allstock_VAE achieved higher accuracy rates than those using individual stock features.It was also concluded that the results obtained with the VAE-reduced stock features were similar to those obtained by own stock features.
基金Sponsored by the National High Technology Research and Development Program of China (Grant No.2002AA721063).
文摘This paper investigates the characteristics of a non-stationary time series, which exists in mechanical fault diagnosis. Combining the characteristics with predictive efficiency, the limitation of the ARIMA model prediction method is analyzed. This model often is applied in the prediction of a non-stationary times series in present. Thus, a wavelet prediction method is introduced to solve non-stationary problems. The Mallat method, often used in signal processing, results form the decimation or the retention of one out of every two samples. Its advantage is that just enough information is kept to allow the exact reconstruction of the input series, but the disadvantage is a time-varying series on line cannot be pursued. Therefore, the authors present another method, à Trous method, which can be applied for recursive prediction in real-time sampling procedure.
基金ItemSponsored by Provincial Natural Science Foundation of Hebei Province of China (E2004000206)
文摘In the strip rolling process, shape control system possesses the characteristics of nonlinearity, strong coupling, time delay and time variation. Based on self adapting Elman dynamic recursion network prediction model, the fuzzy control method was used to control the shape on four-high cold mill. The simulation results showed that the system can be applied to real time on line control of the shape.
基金supported by the Nation Natural Science Foundation of China(NSFC)under Grant No.61462042 and No.61966018.
文摘Traffic flow prediction is an important part of the intelligent transportation system. Accurate multi-step traffic flow prediction plays an important role in improving the operational efficiency of the traffic network. Since traffic flow data has complex spatio-temporal correlation and non-linearity, existing prediction methods are mainly accomplished through a combination of a Graph Convolutional Network (GCN) and a recurrent neural network. The combination strategy has an excellent performance in traffic prediction tasks. However, multi-step prediction error accumulates with the predicted step size. Some scholars use multiple sampling sequences to achieve more accurate prediction results. But it requires high hardware conditions and multiplied training time. Considering the spatiotemporal correlation of traffic flow and influence of external factors, we propose an Attention Based Spatio-Temporal Graph Convolutional Network considering External Factors (ABSTGCN-EF) for multi-step traffic flow prediction. This model models the traffic flow as diffusion on a digraph and extracts the spatial characteristics of traffic flow through GCN. We add meaningful time-slots attention to the encoder-decoder to form an Attention Encoder Network (AEN) to handle temporal correlation. The attention vector is used as a competitive choice to draw the correlation between predicted states and historical states. We considered the impact of three external factors (daytime, weekdays, and traffic accident markers) on the traffic flow prediction tasks. Experiments on two public data sets show that it makes sense to consider external factors. The prediction performance of our ABSTGCN-EF model achieves 7.2%–8.7% higher than the state-of-the-art baselines.
文摘Considering chaotic time series multi-step prediction, multi-step direct prediction model based on partial least squares （PLS） is proposed in this article, where PLS, the method for predicting a set of dependent variables forming a large set of predictors, is used to model the dynamic evolution between the space points and the corresponding future points. The model can eliminate error accumulation with the common single-step local model algorithm~ and refrain from the high multi-collinearity problem in the reconstructed state space with the increase of embedding dimension. Simulation predictions are done on the Mackey-Glass chaotic time series with the model. The satisfying prediction accuracy is obtained and the model efficiency verified. In the experiments, the number of extracted components in PLS is set with cross-validation procedure.
基金supported in part by ZTE Corporation under Grant No.2021420118000065.
文摘A content-aware multi-step prediction control(CAMPC)algorithm is proposed to determine the bitrate of 360-degree videos,aim⁃ing to enhance the quality of experience(QoE)of users and reduce the cost of video content providers(VCP).The CAMPC algorithm first em⁃ploys a neural network to generate the content richness and combines it with the current field of view(FOV)to accurately predict the probability distribution of tiles being viewed.Then,for the tiles in the predicted viewport which directly affect QoE,the CAMPC algorithm utilizes a multi-step prediction for future system states,and accordingly selects the bitrates of multiple subsequent steps,instead of an instantaneous state.Meanwhile,it controls the buffer occupancy to eliminate the impact of prediction errors.We implement CAMPC on players by building a 360-degree video streaming platform and evaluating other advanced adaptive bitrate(ABR)rules through the real network.Experimental results show that CAMPC can save 83.5%of bandwidth resources compared with the scheme that completely transmits the tiles outside the viewport with the Dynamic Adaptive Streaming over HTTP(DASH)protocol.Besides,the proposed method can improve the system utility by 62.7%and 27.6%compared with the DASH official and viewport-based rules,respectively.
基金supported by the National Natural Science Foundation of China(62172089,61972087,62172090).
文摘Virtual machine(VM)consolidation is an effective way to improve resource utilization and reduce energy consumption in cloud data centers.Most existing studies have considered VM consolidation as a bin-packing problem,but the current schemes commonly ignore the long-term relationship between VMs and hosts.In addition,there is a lack of long-term consideration for resource optimization in the VM consolidation,which results in unnecessary VM migration and increased energy consumption.To address these limitations,a VM consolidation method based on multi-step prediction and affinity-aware technique for energy-efficient cloud data centers(MPaAF-VMC)is proposed.The proposed method uses an improved linear regression prediction algorithm to predict the next-moment resource utilization of hosts and VMs,and obtains the stage demand of resources in the future period through multi-step prediction,which is realized by iterative prediction.Then,based on the multi-step prediction,an affinity model between the VM and host is designed using the first-order correlation coefficient and Euclidean distance.During the VM consolidation,the affinity value is used to select the migration VM and placement host.The proposed method is compared with the existing consolidation algorithms on the PlanetLab and Google cluster real workload data using the CloudSim simulation platform.Experimental results show that the proposed method can achieve significant improvement in reducing energy consumption,VM migration costs,and service level agreement(SLA)violations.
文摘This paper studied a fast recursive predictive algorithm used for medical X-ray image compression. This algorithm consists of mathematics model building, fast recursive algorithm deducing, initial value determining, step-size selecting, image compression encoding and original image recovering. The experiment result indicates that this algorithm has not only a higher compression ratio to medical X-ray images compression, but also promotes image compression speed greatly.
基金Item Sponsored by Natural Science Foundation of Hebei Province of China(E2004000206)National Natural Science Foundation of China(50675186)
文摘The rolling force model for cold tandem mill was put forward by using the Elman dynamic recursive network method,based on the actual measured data.Furthermore,a good assumption is put forward,which brings a full universe of discourse self-adjusting factor fuzzy control,closed-loop adjusting,based on error feedback and expertise into a rolling force prediction model,to modify prediction outputs and improve prediction precision and robustness.The simulated results indicate that the method is highly effective and the prediction precision is better than that of the traditional method.Predicted relative error is less than ±4%,so the prediction is high precise for the cold tandem mill.
基金This work was supported by the National Natural Science Foundation of China (No. 60174021, No. 60374037)the Science and Technology Greativeness Foundation of Nankai University
文摘A compound neural network was constructed during the process of identification and multi-step prediction. Under the PID-type long-range predictive cost function, the control signal was calculated based on gradient algorithm. The nonlinear controller’s structure was similar to the conventional PID controller. The parameters of this controller were tuned by using a local recurrent neural network on-line. The controller has a better effect than the conventional PID controller. Simulation study shows the effectiveness and good performance.
基金funded by Prince Sultan University,Riyadh,Saudi Arabia。
文摘For the unforced dynamical non-linear state–space model,a new Q1 and efficient square root extended kernel recursive least square estimation algorithm is developed in this article.The proposed algorithm lends itself towards the parallel implementation as in the FPGA systems.With the help of an ortho-normal triangularization method,which relies on numerically stable givens rotation,matrix inversion causes a computational burden,is reduced.Matrix computation possesses many excellent numerical properties such as singularity,symmetry,skew symmetry,and triangularity is achieved by using this algorithm.The proposed method is validated for the prediction of stationary and non-stationary Mackey–Glass Time Series,along with that a component in the x-direction of the Lorenz Times Series is also predicted to illustrate its usefulness.By the learning curves regarding mean square error(MSE)are witnessed for demonstration with prediction performance of the proposed algorithm from where it’s concluded that the proposed algorithm performs better than EKRLS.This new SREKRLS based design positively offers an innovative era towards non-linear systolic arrays,which is efficient in developing very-large-scale integration(VLSI)applications with non-linear input data.Multiple experiments are carried out to validate the reliability,effectiveness,and applicability of the proposed algorithm and with different noise levels compared to the Extended kernel recursive least-squares(EKRLS)algorithm.
文摘A DRNN (diagonal recurrent neural network) and its RPE (recurrent prediction error) learning algorithm are proposed in this paper .Using of the simple structure of DRNN can reduce the capacity of calculation. The principle of RPE learning algorithm is to adjust weights along the direction of Gauss-Newton. Meanwhile, it is unnecessary to calculate the second local derivative and the inverse matrixes, whose unbiasedness is proved. With application to the extremely short time prediction of large ship pitch, satisfactory results are obtained. Prediction effect of this algorithm is compared with that of auto-regression and periodical diagram method, and comparison results show that the proposed algorithm is feasible.
基金supported by Incheon NationalUniversity Research Grant in 2017.
文摘Video prediction is the problem of generating future frames by exploiting the spatiotemporal correlation from the past frame sequence.It is one of the crucial issues in computer vision and has many real-world applications,mainly focused on predicting future scenarios to avoid undesirable outcomes.However,modeling future image content and object is challenging due to the dynamic evolution and complexity of the scene,such as occlusions,camera movements,delay and illumination.Direct frame synthesis or optical-flow estimation are common approaches used by researchers.However,researchers mainly focused on video prediction using one of the approaches.Both methods have limitations,such as direct frame synthesis,usually face blurry prediction due to complex pixel distributions in the scene,and optical-flow estimation,usually produce artifacts due to large object displacements or obstructions in the clip.In this paper,we constructed a deep neural network Frame Prediction Network(FPNet-OF)with multiplebranch inputs(optical flow and original frame)to predict the future video frame by adaptively fusing the future object-motion with the future frame generator.The key idea is to jointly optimize direct RGB frame synthesis and dense optical flow estimation to generate a superior video prediction network.Using various real-world datasets,we experimentally verify that our proposed framework can produce high-level video frame compared to other state-ofthe-art framework.
文摘A good machine learning model would greatly contribute to an accurate crime prediction. Thus, researchers select advanced models more frequently than basic models. To find out whether advanced models have a prominent advantage, this study focuses shift from obtaining crime prediction to on comparing model performance between these two types of models on crime prediction. In this study, we aimed to predict burglary occurrence in Los Angeles City, and compared a basic model just using prior year burglary occurrence with advanced models including linear regressor and random forest regressor. In addition, American Community Survey data was used to provide neighborhood level socio-economic features. After finishing data preprocessing steps that regularize the dataset, recursive feature elimination was utilized to determine the final features and the parameters of the two advanced models. Finally, to find out the best fit model, three metrics were used to evaluate model performance: R squared, adjusted R squared and mean squared error. The results indicate that linear regressor is the most suitable model among three models applied in the study with a slightly smaller mean squared error than that of basic model, whereas random forest model performed worse than the basic model. With a much more complex learning steps, advanced models did not show prominent advantages, and further research to extend the current study were discussed.
