Two-Way Neural Network Performance PredictionModel Based onKnowledge Evolution and Individual Similarity

Predicting students’academic achievements is an essential issue in education,which can benefit many stakeholders,for instance,students,teachers,managers,etc.Compared with online courses such asMOOCs,students’academicrelateddata in the face-to-face physical teaching environment is usually sparsity,and the sample size is relativelysmall.It makes building models to predict students’performance accurately in such an environment even morechallenging.This paper proposes a Two-WayNeuralNetwork(TWNN)model based on the bidirectional recurrentneural network and graph neural network to predict students’next semester’s course performance using only theirprevious course achievements.Extensive experiments on a real dataset show that our model performs better thanthe baselines in many indicators.

Composition optimization and performance prediction for ultra-stable water-based aerosol based on thermodynamic entropy theory

Water-based aerosol is widely used as an effective strategy in electro-optical countermeasure on the battlefield used to the preponderance of high efficiency,low cost and eco-friendly.Unfortunately,the stability of the water-based aerosol is always unsatisfactory due to the rapid evaporation and sedimentation of the aerosol droplets.Great efforts have been devoted to improve the stability of water-based aerosol by using additives with different composition and proportion.However,the lack of the criterion and principle for screening the effective additives results in excessive experimental time consumption and cost.And the stabilization time of the aerosol is still only 30 min,which could not meet the requirements of the perdurable interference.Herein,to improve the stability of water-based aerosol and optimize the complex formulation efficiently,a theoretical calculation method based on thermodynamic entropy theory is proposed.All the factors that influence the shielding effect,including polyol,stabilizer,propellant,water and cosolvent,are considered within calculation.An ultra-stable water-based aerosol with long duration over 120 min is obtained with the optimal fogging agent composition,providing enough time for fighting the electro-optic weapon.Theoretical design guideline for choosing the additives with high phase transition temperature and low phase transition enthalpy is also proposed,which greatly improves the total entropy change and reduce the absolute entropy change of the aerosol cooling process,and gives rise to an enhanced stability of the water-based aerosol.The theoretical calculation methodology contributes to an abstemious time and space for sieving the water-based aerosol with desirable performance and stability,and provides the powerful guarantee to the homeland security.

A Stacking Machine Learning Model for Student Performance Prediction Based on Class Activities in E-Learning

After the spread of COVID-19,e-learning systems have become crucial tools in educational systems worldwide,spanning all levels of education.This widespread use of e-learning platforms has resulted in the accumulation of vast amounts of valuable data,making it an attractive resource for predicting student performance.In this study,we aimed to predict student performance based on the analysis of data collected from the OULAD and Deeds datasets.The stacking method was employed for modeling in this research.The proposed model utilized weak learners,including nearest neighbor,decision tree,random forest,enhanced gradient,simple Bayes,and logistic regression algorithms.After a trial-and-error process,the logistic regression algorithm was selected as the final learner for the proposed model.The results of experiments with the above algorithms are reported separately for the pass and fail classes.The findings indicate that the accuracy of the proposed model on the OULAD dataset reached 98%.Overall,the proposed method improved accuracy by 4%on the OULAD dataset.

MPDP: A Probabilistic Architecture for Microservice Performance Diagnosis and Prediction

In recent years,container-based cloud virtualization solutions have emerged to mitigate the performance gap between non-virtualized and virtualized physical resources.However,there is a noticeable absence of techniques for predicting microservice performance in current research,which impacts cloud service users’ability to determine when to provision or de-provision microservices.Predicting microservice performance poses challenges due to overheads associated with actions such as variations in processing time caused by resource contention,which potentially leads to user confusion.In this paper,we propose,develop,and validate a probabilistic architecture named Microservice Performance Diagnosis and Prediction(MPDP).MPDP considers various factors such as response time,throughput,CPU usage,and othermetrics to dynamicallymodel interactions betweenmicroservice performance indicators for diagnosis and prediction.Using experimental data fromourmonitoring tool,stakeholders can build various networks for probabilistic analysis ofmicroservice performance diagnosis and prediction and estimate the best microservice resource combination for a given Quality of Service(QoS)level.We generated a dataset of microservices with 2726 records across four benchmarks including CPU,memory,response time,and throughput to demonstrate the efficacy of the proposed MPDP architecture.We validate MPDP and demonstrate its capability to predict microservice performance.We compared various Bayesian networks such as the Noisy-OR Network(NOR),Naive Bayes Network(NBN),and Complex Bayesian Network(CBN),achieving an overall accuracy rate of 89.98%when using CBN.

Performance Degradation Prediction of Proton Exchange Membrane Fuel Cell Based on CEEMDAN-KPCA and DA-GRU Networks

In order to improve the performance degradation prediction accuracy of proton exchange membrane fuel cell(PEMFC),a fusion prediction method(CKDG)based on adaptive noise complete ensemble empirical mode decomposition(CEEMDAN),kernel principal component analysis(KPCA)and dual attention mechanism gated recurrent unit neural network(DA-GRU)was proposed.CEEMDAN and KPCA were used to extract the input feature data sequence,reduce the influence of random factors,and capture essential feature components to reduce the model complexity.The DA-GRU network helps to learn the feature mapping relationship of data in long time series and predict the changing trend of performance degradation data more accurately.The actual aging experimental data verify the performance of the CKDG method.The results show that under the steady-state condition of 20%training data prediction,the CKDA method can reduce the root mean square error(RMSE)by 52.7%and 34.6%,respectively,compared with the traditional LSTM and GRU neural networks.Compared with the simple DA-GRU network,RMSE is reduced by 15%,and the degree of over-fitting is reduced,which has higher accuracy.It also shows excellent prediction performance under the dynamic condition data set and has good universality.

Validation and performance of three scoring systems for predicting primary non-function and early allograft failure after liver transplantation

Background: Primary non-function(PNF) and early allograft failure(EAF) after liver transplantation(LT) seriously affect patient outcomes. In clinical practice, effective prognostic tools for early identifying recipients at high risk of PNF and EAF were urgently needed. Recently, the Model for Early Allograft Function(MEAF), PNF score by King's College(King-PNF) and Balance-and-Risk-Lactate(BAR-Lac) score were developed to assess the risks of PNF and EAF. This study aimed to externally validate and compare the prognostic performance of these three scores for predicting PNF and EAF. Methods: A retrospective study included 720 patients with primary LT between January 2015 and December 2020. MEAF, King-PNF and BAR-Lac scores were compared using receiver operating characteristic(ROC) and the net reclassification improvement(NRI) and integrated discrimination improvement(IDI) analyses. Results: Of all 720 patients, 28(3.9%) developed PNF and 67(9.3%) developed EAF in 3 months. The overall early allograft dysfunction(EAD) rate was 39.0%. The 3-month patient mortality was 8.6% while 1-year graft-failure-free survival was 89.2%. The median MEAF, King-PNF and BAR-Lac scores were 5.0(3.5–6.3),-2.1(-2.6 to-1.2), and 5.0(2.0–11.0), respectively. For predicting PNF, MEAF and King-PNF scores had excellent area under curves(AUCs) of 0.872 and 0.891, superior to BAR-Lac(AUC = 0.830). The NRI and IDI analyses confirmed that King-PNF score had the best performance in predicting PNF while MEAF served as a better predictor of EAD. The EAF risk curve and 1-year graft-failure-free survival curve showed that King-PNF was superior to MEAF and BAR-Lac scores for stratifying the risk of EAF. Conclusions: MEAF, King-PNF and BAR-Lac were validated as practical and effective risk assessment tools of PNF. King-PNF score outperformed MEAF and BAR-Lac in predicting PNF and EAF within 6 months. BAR-Lac score had a huge advantage in the prediction for PNF without post-transplant variables. Proper use of these scores will help early identify PNF, standardize grading of EAF and reasonably select clinical endpoints in relative studies.

Tunnelling performance prediction of cantilever boring machine in sedimentary hard-rock tunnel using deep belief network

Evaluating the adaptability of cantilever boring machine(CBM) through in-depth excavation and analysis of tunnel excavation data and rock mass parameters is the premise of mechanical design and efficient excavation in the field of underground space engineering.This paper presented a case study of tunnelling performance prediction method of CBM in sedimentary hard-rock tunnel of Karst landform type by using tunneling data and surrounding rock parameters.The uniaxial compressive strength(UCS),rock integrity factor(Kv),basic quality index([BQ]),rock quality index RQD,brazilian tensile strength(BTS) and brittleness index(BI) were introduced to construct a performance prediction database based on the hard-rock tunnel of Guiyang Metro Line 1 and Line 3,and then established the performance prediction model of cantilever boring machine.Then the deep belief network(DBN) was introduced into the performance prediction model,and the reliability of performance prediction model was verified by combining with engineering data.The study showed that the influence degree of surrounding rock parameters on the tunneling performance of the cantilever boring machine is UCS > [BQ] > BTS >RQD > Kv > BI.The performance prediction model shows that the instantaneous cutting rate(ICR) has a good correlation with the surrounding rock parameters,and the predicting model accuracy is related to the reliability of construction data.The prediction of limestone and dolomite sections of Line 3 based on the DBN performance prediction model shows that the measured ICR and predicted ICR is consistent and the built performance prediction model is reliable.The research results have theoretical reference significance for the applicability analysis and mechanical selection of cantilever boring machine for hard rock tunnel.

Evaluation and prediction of earth pressure balance shield performance in complex rock strata:A case study in Dalian,China

This research explores the potential for the evaluation and prediction of earth pressure balance shield performance based on a gray system model.The research focuses on a shield tunnel excavated for Metro Line 2 in Dalian,China.Due to the large error between the initial geological exploration data and real strata,the project construction is extremely difficult.In view of the current situation regarding the project,a quantitative method for evaluating the tunneling efficiency was proposed using cutterhead rotation(R),advance speed(S),total thrust(F)and torque(T).A total of 80 datasets with three input parameters and one output variable(F or T)were collected from this project,and a prediction framework based gray system model was established.Based on the prediction model,five prediction schemes were set up.Through error analysis,the optimal prediction scheme was obtained from the five schemes.The parametric investigation performed indicates that the relationships between F and the three input variables in the gray system model harmonize with the theoretical explanation.The case shows that the shield tunneling performance and efficiency are improved by the tunneling parameter prediction model based on the gray system model.

Genomic prediction of yield performance among single-cross maize hybrids using a partial diallel cross design

Genomic prediction(GP)in plant breeding has the potential to predict and identify the best-performing hybrids based on the genotypes of their parental lines.In a GP experiment,34 elite inbred lines were selected to make 285 single-cross hybrids in a partial-diallel cross design.These lines represented a mini-core collection of Chinese maize germplasm and comprised 18 inbred lines from the Stiff Stalk heterotic group and 16 inbred lines from the Non-Stiff Stalk heterotic group.The parents were genotyped by sequencing and the 285 hybrids were phenotyped for nine yield and yield-related traits at two locations in the summer sowing area(SUS)and three locations in the spring sowing area(SPS)in the main maizeproducing regions of China.Multiple GP models were employed to assess the accuracy of trait prediction in the hybrids.By ten-fold cross-validation,the prediction accuracies of yield performance of the hybrids estimated by the genomic best linear unbiased prediction(GBLUP)model in SUS and SPS were 0.51 and 0.46,respectively.The prediction accuracies of the remaining yield-related traits estimated with GBLUP ranged from 0.49 to 0.86 and from 0.53 to 0.89 in SUS and SPS,respectively.When additive,dominance,epistasis effects,genotype-by-environment interaction,and multi-trait effects were incorporated into the prediction model,the prediction accuracy of hybrid yield performance was improved.The ratio of training to testing population and size of training population optimal for yield prediction were determined.Multiple prediction models can improve prediction accuracy in hybrid breeding.

Performance Evaluation of Deep Dense Layer Neural Network for Diabetes Prediction

Diabetes is one of the fastest-growing human diseases worldwide and poses a significant threat to the population’s longer lives.Early prediction of diabetes is crucial to taking precautionary steps to avoid or delay its onset.In this study,we proposed a Deep Dense Layer Neural Network(DDLNN)for diabetes prediction using a dataset with 768 instances and nine variables.We also applied a combination of classical machine learning(ML)algorithms and ensemble learning algorithms for the effective prediction of the disease.The classical ML algorithms used were Support Vector Machine(SVM),Logistic Regression(LR),Decision Tree(DT),K-Nearest Neighbor(KNN),and Naïve Bayes(NB).We also constructed ensemble models such as bagging(Random Forest)and boosting like AdaBoost and Extreme Gradient Boosting(XGBoost)to evaluate the performance of prediction models.The proposed DDLNN model and ensemble learning models were trained and tested using hyperparameter tuning and K-Fold cross-validation to determine the best parameters for predicting the disease.The combined ML models used majority voting to select the best outcomes among the models.The efficacy of the proposed and other models was evaluated for effective diabetes prediction.The investigation concluded that the proposed model,after hyperparameter tuning,outperformed other learning models with an accuracy of 84.42%,a precision of 85.12%,a recall rate of 65.40%,and a specificity of 94.11%.

Building up a general selection strategy and catalytic performance prediction expressions of heteronuclear double-atom catalysts for N_(2)reduction

The severe environmental problems and the demand for energy urgently require electrocatalysis to replace Haber-Bosch for the nitrogen reduction reaction(NRR).The descriptors and important properties of single-atom and homonuclear double-atom catalysts have been preliminarily explored,but the relationship between the inherent properties and catalytic activity of heteronuclear double-atom catalysts with better performance remains unclear.Therefore,it is very significant to explore the prediction expressions of catalytic activity of heteronuclear double-atom catalysts based on their inherent properties and find the rule for selecting catalytic centers.Herein,by summarizing the free energy for the key steps of NRR on 55 catalysts calculated through the first-principle,the expressions of predicting the free energy and the corresponding descriptors are deduced by the machine learning,and the strategy for selecting the appropriate catalytic center is proposed.The selection strategy for the central atom of heteronuclear double-atom catalysts is that the atomic number of central B atom should be between group VB and VIIIB,and the electron difference between central A atom and B atom should be large enough,and the selectivity of NRR or hydrogen evolution reaction(HER)could be calculated through the prediction formula.Moreover,five catalysts are screened to have low limiting potential and excellent selectivity,and are further analyzed by electron transfer.This work explores the relationship between the inherent properties of heteronuclear double-atom catalysts and the catalytic activity,and puts forward the rules for selecting the heteronuclear double-atom catalytic center,which has guiding significance for the experiment.

Data-Driven Probabilistic S for Batsman Performance Prediction in a Cricket Match

Batsmen are the backbone of any cricket team and their selection is very critical to the team’s success.A good batsman not only scores run but also provides stability to the team’s innings.The most important factor in selecting a batsman is their ability to score runs.It is a generally accepted notion that the future performance of a batsman can be predicted by observing and analyzing their past record.This hypothesis is based on the fact that a player’s batting aver-age is generally considered to be a good indicator of their future performance.We proposed a data-driven probabilistic system for batsman performance prediction in the game of cricket.It captures the dependencies between the runs scored by a batsman in consecutive balls.The system is evaluated using a dataset extracted from the Cricinfo website.The system is based on a Hidden Markov model(HMM).HMM is used to generate the prediction model to foresee players’upcoming performances.The first-order Markov chain assumes that the probabil-ity of a batsman scoring runs in the next ball is only dependent on how many runs he scored in the current ball.We use a data-driven approach to learn the para-meters of the HMM from data.A probabilistic matrix is made that predicts what scores the batter can do on the upcoming balls.The results show that the system can accurately predict the runs scored by a batsman in a ball.

Deep Learning for Multivariate Prediction of Building Energy Performance of Residential Buildings

In the quest to minimize energy waste,the energy performance of buildings(EPB)has been a focus because building appliances,such as heating,ventilation,and air conditioning,consume the highest energy.Therefore,effective design and planning for estimating heating load(HL)and cooling load(CL)for energy saving have become paramount.In this vein,efforts have been made to predict the HL and CL using a univariate approach.However,this approach necessitates two models for learning HL and CL,requiring more computational time.Moreover,the one-dimensional(1D)convolutional neural network(CNN)has gained popularity due to its nominal computa-tional complexity,high performance,and low-cost hardware requirement.In this paper,we formulate the prediction as a multivariate regression problem in which the HL and CL are simultaneously predicted using the 1D CNN.Considering the building shape characteristics,one kernel size is adopted to create the receptive fields of the 1D CNN to extract the feature maps,a dense layer to interpret the maps,and an output layer with two neurons to predict the two real-valued responses,HL and CL.As the 1D data are not affected by excessive parameters,the pooling layer is not applied in this implementation.Besides,the use of pooling has been questioned by recent studies.The performance of the proposed model displays a comparative advantage over existing models in terms of the mean squared error(MSE).Thus,the proposed model is effective for EPB prediction because it reduces computational time and significantly lowers the MSE.

Accurate Machine Learning Predictions of Sci-Fi Film Performance

A groundbreaking method is introduced to leverage machine learn-ing algorithms to revolutionize the prediction of success rates for science fiction films.In the captivating world of the film industry,extensive research and accurate forecasting are vital to anticipating a movie’s triumph prior to its debut.Our study aims to harness the power of available data to estimate a film’s early success rate.With the vast resources offered by the internet,we can access a plethora of movie-related information,including actors,directors,critic reviews,user reviews,ratings,writers,budgets,genres,Facebook likes,YouTube views for movie trailers,and Twitter followers.The first few weeks of a film’s release are crucial in determining its fate,and online reviews and film evaluations profoundly impact its opening-week earnings.Hence,our research employs advanced supervised machine learning techniques to predict a film’s triumph.The Internet Movie Database(IMDb)is a comprehensive data repository for nearly all movies.A robust predictive classification approach is developed by employing various machine learning algorithms,such as fine,medium,coarse,cosine,cubic,and weighted KNN.To determine the best model,the performance of each feature was evaluated based on composite metrics.Moreover,the significant influences of social media platforms were recognized including Twitter,Instagram,and Facebook on shaping individuals’opinions.A hybrid success rating prediction model is obtained by integrating the proposed prediction models with sentiment analysis from available platforms.The findings of this study demonstrate that the chosen algorithms offer more precise estimations,faster execution times,and higher accuracy rates when compared to previous research.By integrating the features of existing prediction models and social media sentiment analysis models,our proposed approach provides a remarkably accurate prediction of a movie’s success.This breakthrough can help movie producers and marketers anticipate a film’s triumph before its release,allowing them to tailor their promotional activities accordingly.Furthermore,the adopted research lays the foundation for developing even more accurate prediction models,considering the ever-increasing significance of social media platforms in shaping individ-uals’opinions.In conclusion,this study showcases the immense potential of machine learning algorithms in predicting the success rate of science fiction films,opening new avenues for the film industry.

Characterizing uncertainty in pavement performance prediction

Taking variability and uncertainty involved in performance prediction into account, in order to make the prediction reliable and meaningful, a distribution-based method is developed to predict future PSI. This method, which is based on the AASHTO pavement performance model, treats predictor variables as random variables with certain probability distributions and obtains the distribution of future PSI through the method of Monte-Carlo simulation. A computer program PERFORM using Monte Carlo simulation is developed to implement the numerical computation. Simulation results based on pavement and traffic parameters show that traffic, surface layer material property, and initial pavement performance are the most significant factors affecting pavement performance. Once the distribution of future PSI is determined, statistics such as the mean and the variance of future PSI are readily available.

Prediction model for corrosion rate of low-alloy steels under atmospheric conditions using machine learning algorithms

This work constructed a machine learning(ML)model to predict the atmospheric corrosion rate of low-alloy steels(LAS).The material properties of LAS,environmental factors,and exposure time were used as the input,while the corrosion rate as the output.6 dif-ferent ML algorithms were used to construct the proposed model.Through optimization and filtering,the eXtreme gradient boosting(XG-Boost)model exhibited good corrosion rate prediction accuracy.The features of material properties were then transformed into atomic and physical features using the proposed property transformation approach,and the dominant descriptors that affected the corrosion rate were filtered using the recursive feature elimination(RFE)as well as XGBoost methods.The established ML models exhibited better predic-tion performance and generalization ability via property transformation descriptors.In addition,the SHapley additive exPlanations(SHAP)method was applied to analyze the relationship between the descriptors and corrosion rate.The results showed that the property transformation model could effectively help with analyzing the corrosion behavior,thereby significantly improving the generalization ability of corrosion rate prediction models.

A Physics-informed Deep-learning Intensity Prediction Scheme for Tropical Cyclones over the Western North Pacific

Accurate prediction of tropical cyclone(TC)intensity is challenging due to the complex physical processes involved.Here,we introduce a new TC intensity prediction scheme for the western North Pacific(WNP)based on a time-dependent theory of TC intensification,termed the energetically based dynamical system(EBDS)model,together with the use of a long short-term memory(LSTM)neural network.In time-dependent theory,TC intensity change is controlled by both the internal dynamics of the TC system and various environmental factors,expressed as environmental dynamical efficiency.The LSTM neural network is used to predict the environmental dynamical efficiency in the EBDS model trained using besttrack TC data and global reanalysis data during 1982–2017.The transfer learning and ensemble methods are used to retrain the scheme using the environmental factors predicted by the Global Forecast System(GFS)of the National Centers for Environmental Prediction during 2017–21.The predicted environmental dynamical efficiency is finally iterated into the EBDS equations to predict TC intensity.The new scheme is evaluated for TC intensity prediction using both reanalysis data and the GFS prediction data.The intensity prediction by the new scheme shows better skill than the official prediction from the China Meteorological Administration(CMA)and those by other state-of-art statistical and dynamical forecast systems,except for the 72-h forecast.Particularly at the longer lead times of 96 h and 120 h,the new scheme has smaller forecast errors,with a more than 30%improvement over the official forecasts.

ST-LSTM-SA:A New Ocean Sound Velocity Field Prediction Model Based on Deep Learning

The scarcity of in-situ ocean observations poses a challenge for real-time information acquisition in the ocean.Among the crucial hydroacoustic environmental parameters,ocean sound velocity exhibits significant spatial and temporal variability and it is highly relevant to oceanic research.In this study,we propose a new data-driven approach,leveraging deep learning techniques,for the prediction of sound velocity fields(SVFs).Our novel spatiotemporal prediction model,STLSTM-SA,combines Spatiotemporal Long Short-Term Memory(ST-LSTM) with a self-attention mechanism to enable accurate and real-time prediction of SVFs.To circumvent the limited amount of observational data,we employ transfer learning by first training the model using reanalysis datasets,followed by fine-tuning it using in-situ analysis data to obtain the final prediction model.By utilizing the historical 12-month SVFs as input,our model predicts the SVFs for the subsequent three months.We compare the performance of five models:Artificial Neural Networks(ANN),Long ShortTerm Memory(LSTM),Convolutional LSTM(ConvLSTM),ST-LSTM,and our proposed ST-LSTM-SA model in a test experiment spanning 2019 to 2022.Our results demonstrate that the ST-LSTM-SA model significantly improves the prediction accuracy and stability of sound velocity in both temporal and spatial dimensions.The ST-LSTM-SA model not only accurately predicts the ocean sound velocity field(SVF),but also provides valuable insights for spatiotemporal prediction of other oceanic environmental variables.

Time series prediction of reservoir bank landslide failure probability considering the spatial variability of soil properties

Historically,landslides have been the primary type of geological disaster worldwide.Generally,the stability of reservoir banks is primarily affected by rainfall and reservoir water level fluctuations.Moreover,the stability of reservoir banks changes with the long-term dynamics of external disastercausing factors.Thus,assessing the time-varying reliability of reservoir landslides remains a challenge.In this paper,a machine learning(ML)based approach is proposed to analyze the long-term reliability of reservoir bank landslides in spatially variable soils through time series prediction.This study systematically investigated the prediction performances of three ML algorithms,i.e.multilayer perceptron(MLP),convolutional neural network(CNN),and long short-term memory(LSTM).Additionally,the effects of the data quantity and data ratio on the predictive power of deep learning models are considered.The results show that all three ML models can accurately depict the changes in the time-varying failure probability of reservoir landslides.The CNN model outperforms both the MLP and LSTM models in predicting the failure probability.Furthermore,selecting the right data ratio can improve the prediction accuracy of the failure probability obtained by ML models.

Assessments of Data-Driven Deep Learning Models on One-Month Predictions of Pan-Arctic Sea Ice Thickness

In recent years,deep learning methods have gradually been applied to prediction tasks related to Arctic sea ice concentration,but relatively little research has been conducted for larger spatial and temporal scales,mainly due to the limited time coverage of observations and reanalysis data.Meanwhile,deep learning predictions of sea ice thickness(SIT)have yet to receive ample attention.In this study,two data-driven deep learning(DL)models are built based on the ConvLSTM and fully convolutional U-net(FC-Unet)algorithms and trained using CMIP6 historical simulations for transfer learning and fine-tuned using reanalysis/observations.These models enable monthly predictions of Arctic SIT without considering the complex physical processes involved.Through comprehensive assessments of prediction skills by season and region,the results suggest that using a broader set of CMIP6 data for transfer learning,as well as incorporating multiple climate variables as predictors,contribute to better prediction results,although both DL models can effectively predict the spatiotemporal features of SIT anomalies.Regarding the predicted SIT anomalies of the FC-Unet model,the spatial correlations with reanalysis reach an average level of 89%over all months,while the temporal anomaly correlation coefficients are close to unity in most cases.The models also demonstrate robust performances in predicting SIT and SIE during extreme events.The effectiveness and reliability of the proposed deep transfer learning models in predicting Arctic SIT can facilitate more accurate pan-Arctic predictions,aiding climate change research and real-time business applications.