最新刊期
卷 21 ， 期 9 ， 2022
- 摘要：In view of the problems that traditional matrix factorization algorithms cannot mine deep hidden information and insufficient utilization of user and item comments， propose a deep collaborative recommendation model based on attention mechanism. Firstly， the attention mechanism is introduced to weight the review text， and the parallel convolutional neural network is used to extract user reviews and item review features respectively， At the same time， the rating matrix is input into the multi-layer perceptron to get the user implicit representation and the item implicit representation. And then integeter the user features and item features. Finally，a factorization machine and a deep neural network are used to extract linear and non-linear features respectively for scoring prediction. The RMSE reached 0.83 on the open data set of Amazon. The RMSE was readuced by 14.0%，11.2%，9.8%，7.7%，3.9% respectively，compared with five groups of models. Experiments showed that the proposed model can improve the recommendation effect effectively.关键词：rating matrix;review text;convolutional neural network;attention mechanism
- 435
- |
- 98
- |
- 0
发布时间：2022-09-21 -
Automatic Punctuation of Ancient Books Based on BERT +BiLSTM+CRF Model and New Preprocessing Method AI导读
摘要：Compared with modern literatures， ancient Chinese not only have huge differences in terms of words and grammar， but also lack punctuation，which is difficult to understand. Punctuation of ancient texts manually requires both high professional knowledge and a certain understanding of the history and culture of the time. In order to improve the accuracy of automatic punctuation of ancient texts， this article tries to combine the deep language model Bidirectional Encoder Representation from Transformers （BERT）， the bidirectional long short memory network and the conditional random field model （BiLSTM+CRF） for automatic punctuation of ancient books， and proposes a new data preprocessing method. The improved model can reach about 85% in the automatic punctuation of ancient texts， which is about 8% higher than the previous common methods. At the same time， the model also shows good generalization performance， even on ancient texts that have never been trained and predicted， the prediction result can reach about 78%. The experimental results show that the improved model and the new preprocessing method can not only better learn the semantic information and contextual information of ancient texts， but also learn the standard information of tags.关键词：ancient Chinese book;automatic punctuation adding;BERT model;conditional random field- 238
- |
- 40
- |
- 0
发布时间：2022-09-21 - 摘要：In order to solve the problem of sparse item scoring matrix， a fuzzy feature combining item features and user ratings is proposed. First， Gaussian like fuzzy numbers are used to describe the membership degree of the category to which the project belongs， then trapezoidal fuzzy numbers are used to represent the user's preference for the project， and a user project category preference matrix is constructed. Finally， a method based on the fuzziness of project characteristics and user interest is constructed to calculate the recommendation score. The experimental results on movielens 100k dataset show that when n is 1~300， the average recommendation accuracy of this algorithm is improved by 39.97% and 5.74% respectively compared with the user based and item based collaborative filtering recommendation algorithms without losing the top-N recommendation recall rate. It effectively solves the problem of data sparsity， and can recommend items of interest to users in the case of less historical behavior data.关键词：collaborative filtering recommendation algorithm;fuzzy numbers;user interest model;recommendation trust score;user preference matrix
- 153
- |
- 28
- |
- 0
发布时间：2022-09-21 - 摘要：Traffic light detection， a core technology in automatic driving， is directly related to the safety of intelligent vehicles. An improved YOLOv5 traffic light detection method is proposed to handle small-size-light recognition and complex-environment anti-interference. In this paper， composite data augmentation is used to increase model input capacity and complexity. Multi-scale training is designed to replace fixed scale training so that the learning ability of the model can be balanced. The proposed also fuses the samples at 4， 8， 16 and 32 times respectively to build a multi-scale detection layer. In order to improve the feature fusion， remote-jump links are introduced to convey different levels of information， which directly improves the detecting precision of small targets. At last， experiments show that the fastest detection speed of improved YOLOv5 can reach 9.5ms， and mAP can reach 99.8%， which is 17% higher than that of YOLOv5， and 6.5% more mAP is added on Bosch dataset， achieving a real-time and high-precision detection of traffic lights.关键词：object detection;YOLOv5;traffic light detection;multi-scale feature;real-time detection
- 290
- |
- 40
- |
- 0
发布时间：2022-09-21 - 摘要：With the popularity of social credit consumption， the number of defaulting users to be collected is gradually increasing. To solve this problem， this paper improves the natural language understanding module in the development framework of Rasa dialogue system， significantly improves the accuracy of the natural language understanding module in the intelligent collection dialogue system， and evaluates the performance of the final training model. In the algorithm model proposed in this paper， the user input text is processed by the domain oriented pre training language model and input to the intention recognition and slot recognition module. In the model proposed in this paper， the recursive convolution neural network model is used for intention recognition， and the gated loop unit and conditional random field model are used to complete the task of slot recognition. Finally， by comparing with the original pre training language model and other mainstream algorithms， F1 values in intention recognition and slot recognition reach 95.75% and 95.88% respectively， which are better than other mainstream algorithms， which verifies the feasibility of the algorithm proposed in this paper.关键词：intelligent collection system;Rasa;natural language understanding;pre training language model
- 54
- |
- 8
- |
- 0
发布时间：2022-09-21 - 摘要：Target detection in low light background is one of the main tasks of night patrol robot for airport terminal. In order to sufficiently extract and make use of the motion information in the dynamic scene video data， improve the detection effect of the detection method in the face of the actual complex situations such as the human body is blocked and incomplete human target at the edge of the image， and avoid the ambiguity caused by the incomplete detection results of a single frame， a salient human body detection method incorporating motion information under thermal imaging video of the airport terminal is proposed. First， the foreground and background of image data are separated preliminarily by a background separation model. Then， the background area misjudged by camera motion is further separated by a feature point track clustering and motion estimation affine model. Finally， the motion target detection result is fused with the results of single frame detection algorithm as motion information.It can effectively extract motion information from dynamic scene video data. Experiments on four datasets show that this method can effectively improve the accuracy of single-frame detection and avoid the occurrence of incomplete detection.关键词：thermal imaging;target detection;motion information;dynamic scene;salient human body detection
- 30
- |
- 7
- |
- 0
发布时间：2022-09-21 - 摘要：Aiming at the problems of slow manual detection of high-voltage substation circuit breakers and large errors， this paper proposes a research method based on YOLOV4 improved substation circuit breaker switching status identification. In view of the complex background of the power system and the difficulty of identifying the opening and closing status of the circuit breaker， this paper adds the channel attention mechanism to focus on the salient features of the target and ignore the non-target areas. Then use the path aggregation network to effectively extract the target features. Aiming at the unity of data samples， this paper proposes the SE-YOLOv4 algorithm to add data enhancement technology to it to improve the generalization ability of the model， and the algorithm network has better robustness. In the data set used in this paper， the accuracy of the experimental results is 97%， the recall rate is 73.45%， the F1 is 0.84， and the average accuracy is 79.00%， which is 2.45% higher than the average accuracy of the original algorithm. Therefore， the detection method based on deep learning can detect the target quickly and efficiently， and avoid the problems of manual detection.关键词：circuit breaker;deep learning;attention mechanism;path aggregation network
- 45
- |
- 9
- |
- 0
发布时间：2022-09-21 - 摘要：Multidimensional time series forecasting has always been an important research direction for scholars. In recent years， long short-term memory network （LSTM） has a huge advantage in processing the nonlinear part of time series. However， a single predictive model cannot take into account the linear and nonlinear characteristics of the data at the same time. To solve this problem， the time series matrix factorization technology TRMF （Temporal Regularized Matrix Factorization） is introduced to process the linear main part of the multivariate time series. After the residual of the training data is calculated， it is input into the LSTM model for nonlinear fitting. Then， the test data is substituted into the trained TRMF-LSTM model， and the linear subject predicted by the model is added to the residual error to obtain the combined prediction value. Select the Shanghai and Shenzhen 300， Shanghai Stock Exchange Index two stock indexes， SANY Heavy Industry， China Life， Agricultural Bank， Muyuan， Midea Group， Longji Stocks 6 stocks， a total of 8 stock price time series for forecasting， using LSTM， Transformer， SVR as Compare the models， and select two evaluation indicators， MAPE and RMSE. The experimental results show that， compared with the comparison model， the minimum values of MAPE and RMSE all fall in the TRMF-LSTM combined prediction model， which fully verifies the effectiveness of the model.关键词：multidimensional time series;time series matrix decomposition;LSTM;combination forecast
- 35
- |
- 11
- |
- 0
发布时间：2022-09-21 - 摘要：Recommender system technology can help users find items of interest quickly and save them time. It can also help businesses cost less. This paper presents a collaborative filtering recommender algorithm based on the similarity of user preferences in order to improve the prediction accuracy and scalability of existing recommender algorithms. By analyzing users' emotional preferences for items， rating predictions and recommendations are made based on the similarity between users， which is calculated by the like index and dislike index. Multiple experiments on the MovieLens dataset show that the method proposed in this paper improves the accuracy of the prediction results. Compared with several other classic algorithms， the algorithm proposed in this paper reduces the number of neighbors required for optimal prediction and the time complexity is lower. It also has good performance and scalability at the same time.关键词：recommender systems;collaborative filtering;rating prediction;user preference
- 30
- |
- 7
- |
- 0
发布时间：2022-09-21 - 摘要：In order to reduce the gap between the application front-end and the system structure model， an approach of mapping the interactive flow modeling language （IFML） model into the service content model （SCM） model is proposed. Firstly， the meta model of the IFML and SCM are analyzed， and the mapping rules from the IFML model elements to SCM model elements are designed. Secondly， the mapping rules are described by Query View Transformation Operational （QVTo）， and are implemented by the QVTo engine on the IBM Rational Software Architect （RSA） platform to complete the automatic mapping. Finally， the online shopping system is applied to demonstrate the executing results of the model mapping， and the consistency between the IFML model and the SCM model is validated by formal semantics. The experimental results show that there is a strong consistency between the elements such as view container， action and event in the application front-end and the system structure. Thus， the automatic mapping from IFML model to SCM model can make the content model design of the business system more in line with the user requirements， and effectively improve the quality and efficiency of software development.关键词：interactive flow modeling language;model mapping;service content model;query view transformation operational
- 32
- |
- 48
- |
- 0
发布时间：2022-09-21 - 摘要：In order to improve the shortcomings of traditional soft-sensing models that are easy to fall into local optimality and low accuracy， a kind of MLSSVR with mixed kernel parameters is used for soft measurement modeling， and the improved Fireworks Competition Fireworks Algorithm Optimization is used to optimize its parameters.Taking Marine lysozyme as the research object， the simulation results show that compared with FWA-MLSSVR and IFWA-LSSVR， the RMS error calculation of this model is reduced by 0.3439 and 0.7462， respectively. The mixed core parameter method effectively avoids local optimization， IFWA improves the prediction accuracy of the model. The prediction model satisfies the design requirements well and has high application value.关键词：marine lysozyme;least squares support vector machine;fireworks algorithm;soft sensing
- 15
- |
- 5
- |
- 0
发布时间：2022-09-21 - 摘要：In order to eliminate radio frequency interference （RFI） and reduce the loss of astronomical signal， a radio frequency interference elimination method based on singular value optimization segmentation is proposed. Firstly， the radio observation signal is decomposed into subbands to construct the observation signal matrix； Then singular value decomposition （SVD） is performed on the matrix to find the singular value corresponding to the interference signal， and the singular value is optimally segmented according to the principle of maximum output signal-to-noise ratio to obtain the optimal segmentation ratio and the optimized singular value component； Finally， RFI signal is removed. The simulation and experimental results show that the optimal segmentation ratio K/N = 5 / 20 obtained by this method for the radio observation signals of the same radio source in different time periods is universal. The output signal-to-noise ratio gain after RFI signal suppression using the optimal segmentation ratio is obvious， which can reach 1.2 ~ 1.4. It can effectively suppress RFI signal， reduce astronomical signal loss， and obtain the maximum output signal-to-noise ratio.The proposed method provides a new idea for eliminating radio frequency interference.关键词：radio frequency interference;interference mitigation;singular value decomposition;optimal segmentation
- 18
- |
- 5
- |
- 0
发布时间：2022-09-21 - 摘要：In order to better analyze the lane changing behavior in the interweaving area of urban tunnel entrances and exits， the characteristics related to the lane changing behavior in the interweaving area of urban tunnel entrances and exits are studied and modeled. Firstly， based on the video data of real traffic flow in the interweaving area taken on site， we extracted the data of vehicles with lane change behavior， analyzed the distribution law of lane change behavior in four aspects： speed and acceleration， lane change point and lane change duration distance， main and auxiliary road difference， and whether there is a conflict， and used multi-classification logit analysis and neural network modeling methods to build the prediction model of lane change in the interweaving area respectively， and used simulation experiments to the accuracy and feasibility of the models were compared and tested. The results show that all the models can reasonably describe the actual traffic flow in the intersection area， and the predicted values of the neural network model are more suitable to the measured values and more consistent with the behavioral characteristics of drivers， which is more applicable.关键词：traffic flow;vehicle behavior characteristics;urban tunnel interweaving area;lane change model;traffic simulation
- 18
- |
- 5
- |
- 0
发布时间：2022-09-21 - 摘要：Because of the anonymity and non-contractual nature of mobile crowdsourcing test， it is difficult to estimate the credibility of crowdsourced testers and guarantee the quality of test results. To solve this problem， this paper proposes a compound method to calculate the credibility of crowdsourced testers. On the basis of the scoring mechanism， confidence intervals for reliable ratings are calculated to continuously update the individual ratings of the crowdsourced testers and to determine the subjective credibility of the crowdsourced testers. The objective credibility of the crowdsourced testers is determined by using hierarchical analysis to determine the weighting of different indicators on the credibility of the testers. Finally， the validity and correctness of the model are verified by simulation experiments.The proposed method reduces the influence of human subjectivity from both subjective and objective perspectives， while at the same time integrating the influence of different indicators， effectively ensuring the credibility of the crowdsourced testers and the quality of the final test results.关键词：mobile application;crowdsourced testing;reputation;analytic hierarchy process;confidence interval
- 18
- |
- 3
- |
- 0
发布时间：2022-09-21 - 摘要：The simulation effect of bubble motion details has an important influence on the reality and intuitiveness of fluid simulation. The smooth particle hydrodynamics method is used to establish the gas-liquid models respectively， and a simple method is used to realize the simulation of bubble suspension. Aiming at the high time and space complexity of neighborhood search， a search algorithm based on interval division combined with hash table is proposed. By dividing the space area， the divided space is mapped to the hash table， which reduces the search space range. The simulation results show that the above method can effectively reduce the time and space complexity and improve the real-time performance of bubble simulation while ensuring the visual effect of the simulation.关键词：smooth particle;hydrodynamics;neighborhood search;fluid simulation;bubble simulation
- 18
- |
- 4
- |
- 0
发布时间：2022-09-21 -
Design and Implementation of Core Experimental Demonstration Platform of Operating System Course AI导读
摘要：Under the background of "double first-class" construction， in order to reduce the difficulty of operating system core experiment course， an interactive core experiment demonstration platform is proposed. The platform has the functions of parameter setting， function demonstration and so on， and presents the running process of a variety of job scheduling， process scheduling and banker algorithm in visual form. The practice results show that the average score of students' final examination has increased from 71.3 to 80.2， and the completion rate of the experiment has increased from 59% to 89%， which has achieved good teaching results.The platform can significantly improve the teaching quality of operating system courses， and provide a certain reference for colleges and universities to carry out the reform of operating system courses.关键词：operating system teaching;core experiment demonstration platform;scheduling algorithm;banker algorithm- 54
- |
- 21
- |
- 0
发布时间：2022-09-21 - 摘要：Aiming at the problem that most laboratory security learning platforms have changed from traditional PC website to mobile APP， the mobile APP which originally independently develops Android and iOS platform has a huge workload and high operation and maintenance cost. Based on flutter development framework and sqflite database plug-in， a cross platform laboratory security learning platform is designed and developed. The platform uses a set of program code written by Flutter development framework to run on Android and iOS platforms at the same time. The security knowledge of the laboratory that learners need to master has constructed three functional modules： Security learning question base， learning test and problem analysis， so that the learners can really improve the awareness of experimental security and safe operation skills. Practice shows that the development framework of Flutter greatly improves the efficiency of cross-platform development， and the learning platform has stable performance. It can run on terminal devices of Android and iOS platforms with good use effect.关键词：learning platform;Flutter;cross-platform;Laboratory;safety learning
- 45
- |
- 12
- |
- 0
发布时间：2022-09-21 - 摘要：In order to meet the urgent needs of comprehensive integration and flexible customization in the information construction of large enterprises and institutions， and solve the growing integration and management problems between applications with different architectures， development frameworks and operating environments in the information ecology， the strategy analysis and research of user data’s storage， binding， mapping， synchronization and access control are used to implement system integration application in the application support platform which organically combine the decentralized application through user information. The results show that the prototype system can effectively implement the proposed user integration strategy. The implementation of user integration based on Apereo CAS and Apache SHIRO framework can not only implement and verify the integration strategy， but also have engineering application value which brings the advantages of agile development and rapid implementation.关键词：system integration application;user integration strategy;Apereo CAS;Apache SHIRO;application support platform
- 71
- |
- 7
- |
- 0
发布时间：2022-09-21 - 摘要：Now we have entered the era of big data， and the amount of data people are exposed to is growing exponentially. At this time， the performance of MySQL database is facing a great test. Therefore， how to improve the data extraction efficiency in MySQL database and make it bear greater concurrency is an urgent problem to be solved. Taking the locally deployed recommendation system as the background， the performance of the database is extended. Firstly， the performance is optimized in the case of a single database， then the anti concurrency ability of the database is improved by deploying the master-slave replication database， and finally the stress test is carried out. According to the index observation， the average response time of the optimized system under 500， 600 and 700 concurrency is reduced by 44ms， 63ms and 109ms respectively， and the throughput is increased by 123 / s， 194 / s and 238 / s. It is further verified that the optimized system has significantly improved in data extraction and anti concurrency performance of the database.关键词：MySQL database;performance tuning;database master-slave deployment;performance testing
- 43
- |
- 7
- |
- 0
发布时间：2022-09-21 -
Research on the Application of AM on the Human Resource Information Management System Promotion AI导读
摘要：At present， the entity relationship model （ERM） is mostly used to design and implement the database in information management system. Although the entity relation model can describe the entities， attributes and relationship between entities， it neither conducive to store the dynamic entities and relationships nor conducive to works such as data validation. In addition， fat tables in the entity relationship model are not conducive to the upper application modules to update themselves. With the development of database technology， the Anchor Model （AM） is provided. It emphasizes the expression and application of the meta-information time in entities， attributes and relationship between entities. Description ability of all kinds of complex scenes is extended by AM’s mathematic model and UML model. The data of AM description is the collection of Schema， not SQL language. And merger of schema are used to support the upper modules of the integration and incremental development. Therefore， AM complements the weakness of ERM and has a clear engineering value for the design and promotion of the database.关键词：anchor model;entity relationship model;information management system;system promotion;human resource- 20
- |
- 4
- |
- 0
发布时间：2022-09-21 - 摘要：With the continuous development of computer technology， the data in the network presents complex， large， redundant and multidimensional characteristics. To improve the detection performance of deep learning-based intrusion detection system， this paper proposes a Transformer-based intrusion detection method. First， the data are preprocessed to make them meet the input requirements of the neural network， and the SMOTE-GMM algorithm is used to solve the data imbalance problem. Second， the Transformer encoder is used to extract the input features， the attention link between the encoder output and the decoder input is established by the Transformer decoder， and finally the classification is completed by softmax. To evaluate the model detection performance， experimental validation is conducted on the NSL-KDD data set. The experimental results show that the Transformer-based intrusion detection method has significant performance improvement compared with DNN， AIDS， FEEM and other methods， and the classification accuracy reaches 88.2% and the precision rate reaches 89.7%.关键词：intrusion detection;data imbalance;SMOTE-GMM;Transformer;NSL-KDD
- 91
- |
- 16
- |
- 0
发布时间：2022-09-21 - 摘要：Aiming at the problems of large waste of resources and low efficiency of data processing in distributed honeypot， a distributed honeypot deployment strategy based on cooperative game is proposed. The cooperative game and distributed honeypot technology are combined， and the clustering and shunting technology is added for data processing. Considering a variety of influencing factors， the resource allocation is carried out in combination with the demand， On the premise of making full use of system resources， maximize the income of honeypot and optimize the performance of honeypot system. The model and two control groups were tested for 72 hours. The experimental results show that the deployment strategy can maximize the honeypot revenue according to the demand and allocate it. Compared with the single deployment strategy， the experimental results reduce the error rate by 12.29%， the system resource loss rate by 7.5% and the data processing time by 59.22%.关键词：network security;distributed honeypot;cooperative game;Shapley value;clustering algorithm
- 37
- |
- 8
- |
- 0
发布时间：2022-09-21 - 摘要：At present， many software is composed of shared open source components. Once a component has a security problem， it will have a significant impact.In order to understand the safety left shift conditions， the error and cause of the security status of the open source component obtained by the software component analysis （SCA） method. This article combed common scenes in security left shifts in this paper， and build test cases based on this. Use the transformation test method to use the SCA tool for safety testing and statistical analysis. Experimental results show that binary SCA and compiling SCAs， overlapping part of data accounts for no more than 60%. BOM-based SCA is compiled in compilation mode， which is more than 90%， and the source code is not compared to the accuracy recall of components in the compiled mode. At about 40%， the relevant indicators are around 90% for the detection of vulnerabilities. It shows that the second-pieces of binary product detection of the software supply chain and the source code are inadequate， and the accuracy of the security left-phase detection test results in the unable to compile state is insufficient， and the BOM file security test results are effectively reliable.关键词：open source components;software component analysis;safe left shift;vulnerability detection
- 46
- |
- 12
- |
- 0
发布时间：2022-09-21 - 摘要：Accurate segmentation of street scene image plays an important auxiliary role in automatic driving system. Although existing semantic segmentation methods for this scene have achieved some results， there are still some problems such as low segmentation accuracy and large number of parameters. In order to improve the performance of semantic segmentation effectively， an attention semantic segmentation network is proposed by constructing spatial attention module and channel attention module. Firstly， the residual network is used to extract features， and then two kinds of attention modules are used to refine feature maps adaptively from spatial and channel dimensions in parallel， so as to make the network pay more attention to information-rich spatial regions and channels in the process of training and learning， so as to enhance the representation ability of the network. The proposed attention module has the characteristics of simple structure and lightweight， and can carry out end-to-end training with the network. Experimental results on Cityscapes and CamVid data sets show that the proposed attentional semantic segmentation network achieves better segmentation performance with fewer parameters.关键词：street scenes images;semantic segmentation;residual network;spatial attention module;channel attention module
- 52
- |
- 7
- |
- 0
发布时间：2022-09-21 - 摘要：To solve the problem that the ground clutter in Doppler weather radar echo map can only be recognized by naked eye， an algorithm based on convolution neural network model and area overlap gray level symbiosis matrix is presented to identify the specific area where the clutter exists in the radar echo map. Firstly， a convolution neural network model is designed， and the convolution neural network is trained with the dataset to recognize the pitch angle of the radar echo map. Then， the area overlap gray level symbiosis matrix algorithm is used to extract the features of low elevation weather radar echo map， and detect the specific area of ground clutter. Finally， the coordinates of the specific area where there is ground clutter are obtained. Tests show that this design can effectively identify the clutter in the meteorological radar echo map.关键词：meteorological radar echo map;convolutional neural network model;region overlap gray level co-occurrence matrix algorithm;ground clutter
- 29
- |
- 5
- |
- 0
发布时间：2022-09-21 - 摘要：Aiming at the problem of insufficient data set and accuracy in image annotation， an automatic image annotation method （FLTAN） combining reinforcement learning and transfer adversation network was proposed.In feature extraction， the ResNet50 feature parameters trained by ImageNet are transferred to the model and add reinforcement learning to the feature extraction for improving precision.Then the extracted features are input into the adversarial model， and we perform feature clipping/standardization operation in the generator part to further improve the model performance.In addition is further improved， in order to complete the Chinese annotation of the image， machine translation is added at the end of the model ，In order to translate English into Chinese to captial the image.The proposed method shows advantages in data sets Image CLEF-DA and DeepFasion2，and the DeepFazion2 image is dealt into color and white picture to compare with colorful picture.as we expected，the color images show better scores .Firstly， this paper briefly reviews the concepts of reinforcement learning， transfer learning， machine translation and attention mechanism.Secondly， the idea of feature extraction of reinforcement learning mode was analyzed， and the ablation experiment was conducted with DAAN， the method of unreinforcement learning. Finally， the experimental results and the advantages and disadvantages of the model were summarized， and the future development trend was analyzed and prospected.关键词：image class;image caption;machine translation;adversarial network;reinforcement learning;transfer learning
- 34
- |
- 8
- |
- 0
发布时间：2022-09-21 - 摘要：In order to improve the retrieval efficiency of agricultural meteorological popular science video， this paper takes agricultural meteorological popular science video as an example， and proposes a shot detection algorithm based on dual detection model （initial detection and re detection）. In the initial detection stage， each frame image is divided into equal area rectangular rings to highlight the main part of the video image center， and then the adaptive double threshold method is used to segment the video for the first time； The improved sift （scale invariant feature transform） algorithm is used to further modify the initial detection results. The segmentation experiments of three agricultural popular science videos show that the overall algorithm design in this paper can segment agricultural meteorological popular science videos effectively， with an average recall rate of 91.2% and an average precision rate of 92.3%. The algorithm meets the professional and personalized retrieval needs of farmers， improves the utilization of agrometeorological videos， and promotes the development of agricultural economy and meteorological science popularization.关键词：agricultural meteorological science video;equal-area rectangular ring;adaptive double threshold;improved sift algorithm
- 22
- |
- 3
- |
- 0
发布时间：2022-09-21 - 摘要：Blockchain audit is beneficial for the development of the digital economy， so the demand for relevant talents has increased rapidly. Based on the analysis of the current demand for talents of the blockchain audit， this paper explores a talent training model of the blockchain audit specialty by adhering to the OBE education concept. Furthermore， a feasible way of nurturing high-level talents in blockchain audit is investigated from the aspects of setting talent training goals and major positioning， developing faculty strength， building a multi-disciplinary course system for blockchain audit speciality， and improving the evaluation mechanism for both teaching and learning， etc.关键词：blockchain audit;OBE;construction mode;talent training;curriculum system
- 41
- |
- 11
- |
- 0
发布时间：2022-09-21 - 摘要：Traditional digital image processing courses usually adopt teacher-and-syllabus-derived teaching mode， ignoring the differences in course knowledge requirements for different career plans. In view of this deficiency， a student-centered， employment goal-oriented course reform of digital image processing is proposed by setting up diversified teaching goals， adopting a dual-driven teaching system of course syllabus and classic cases and online and offline mixed teaching mode to realize multi-track classified teaching. Students can choose the corresponding learning channel according to their personal career planning to achieve targeted training of core competencies and literacy. The reform of digital image processing course based on students' goal orientation can stimulate the inner driving force of students to the greatest extent and improve their autonomous learning ability， finally improving the teaching quality.关键词：digital image processing;students' goal orientation;course reform;classic cases;online and offline mixed mode
- 26
- |
- 10
- |
- 0
发布时间：2022-09-21 - 摘要：With the requirements of strengthening innovation ability of the students under the background of the new engineering construction， this paper focus on the teaching mode reform of mobile communication course. Mobile communication course undertakes the connection between the basic theoretical knowledge of the courses such as communication principles and a series of communication technologies employed in the 5G mobile communication network， which plays an important role in cultivating the scientific research innovation and practical ability of the students. When teaching the contents of 5G related new mobile communication technology， group discussion methods are employed to study and discuss academic papers associated with the current research topics， which is useful for developing the initiative exploratory learning and the scientific research thinking and innovation ability of the students.关键词：new engineering;mobile communication;5G;teaching contents reform
- 19
- |
- 9
- |
- 0
发布时间：2022-09-21 - 摘要：Aiming at the problems existing in the engineering teaching mode and traditional OBE concept curriculum in Colleges and universities in China. Taking discrete mathematics course as an example， this paper puts forward the teaching goal of combining the requirements of knowledge， ability and literacy. By constructing the teaching reform framework of "knowledge point - teaching requirements - teaching objectives - teaching methods"， the goal refinement method to meet the teaching requirements of knowledge point level is realized. Firstly， through the output oriented improvement mechanism， discrete mathematical knowledge points in selected cases are extracted to establish the correlation between knowledge points. Then， the ability map based on the knowledge point level teaching requirements is constructed to accurately evaluate the diversified teaching effect. Teaching practice has proved that this teaching mode realizes the early warning and assistance to students with learning difficulties， improves the fine achievement effect of teaching objectives， and significantly improves students' autonomous learning ability. The proposed method is expected to provide reference for the curriculum reform of discrete mathematics.关键词：discrete mathematics;teaching goal;refined achievement;knowledge points;self-regulated learning
- 15
- |
- 5
- |
- 0
发布时间：2022-09-21 - 摘要：With the rapid increase in the application of virtual simulation systems in the field of higher education， understanding the application status of virtual simulation experimental teaching projects is of great significance to promoting the information reform of higher education and improving the quality of experimental teaching.The paper takes the 2017-2021 virtual simulation experimental teaching projects opened by the National Virtual Simulation Experimental Teaching Course Sharing Platform （also known as "Virtual Space"） as the research object， and expands detailed analysis on the current construction status and application of the virtual simulation experimental teaching project. In view of the existing problems， we learn from the experience， and further propose optimization strategies to provide practical enlightenment for the future application and development of virtual simulation experiment teaching project.关键词：virtual simulation;higher education;education informatization;experimental teaching
- 21
- |
- 6
- |
- 0
发布时间：2022-09-21 - 摘要：A novel teaching strategy is proposed by providing students with adaptive learning activities based on a revised Bloom taxonomy of learning theory. The adaptive feature is achieved by employing fuzzy weight-based decisions that define students’ knowledge levels and infer learning material based on their scores on domain concepts. The result is that each student receives learning activities according to his level of knowledge， and they are adaptable in number， type and complexity. By comparing the system with two traditional versions of the system， the results show that the system outperforms the traditional version in terms of improving learning outcomes， improving the efficiency of adaptability and RBT use， and the effectiveness of teaching strategies using learning activities.关键词：Revised Bloom Taxonomy （RBT）;computer programming;teaching strategies;fuzzy weights;knowledge level;learning activities;adaptability
- 16
- |
- 9
- |
- 0
发布时间：2022-09-21 - 摘要：Under the background of "double first-class" construction， should focus on cultivating the practical innovation ability of computer major graduate students. However， there are still some problems in the current training work， such as the deviation of training objectives， loose practice management and single evaluation system. Therefore， the continuous improvement model of professional degree training mode based on PDCA is proposed. By cooperating with the practice teaching system of school enterprise tutors and the whole process wide caliber professional degree evaluation index system， the knowledge and ability structure design of practice links is optimized， the quality of practice teaching and professional degree graduate training is improved， and a brand-new evaluation index is constructed. The practice results show that the reform measures are helpful to improve the engineering practice ability of computer major graduate students and improve the quality of postgraduate talent training， so as to provide reference and reference for colleges and universities carrying out the degree education reform of computer major.关键词：professional degree in computer science;continuous improvement;school-enterprise tutor cooperation;evaluation index system
- 11
- |
- 5
- |
- 0
发布时间：2022-09-21 - 摘要：To explore the triggering process of College Students' deep learning behavior in mixed learning environment. By using the design based research （DBR） method， this paper explores the hybrid teaching path to promote the deep learning of college students through practice in the real situation of higher education. The practice results show that teachers' energy and time are the main obstacles to the implementation of hybrid teaching， but the hybrid teaching path based on DBR can help teachers find teaching problems in time and revise the teaching process iteratively， improve students' critical thinking and unity and cooperation ability， promote students' deep learning， and provide reference and reference for colleges and universities to carry out deep learning teaching design.关键词：blended learning environments;deep learning;design-based research;instructional pathways;collaborative learning
- 12
- |
- 6
- |
- 0
发布时间：2022-09-21 - 摘要：To speed up the implementation of teaching innovation and improve the effectiveness of building morality and cultivating people. To explore the deep teaching mode of “student-centered development and competence-oriented” through reflecting and improving the teaching scheme of the machine vision course in the context of the new engineering discipline in the process of teaching practice.Through the study analysis， the key problems are deeply explored. For instance， the difficulty of constructing a knowledge system， the lack of practical training， and the lack of in-depth classroom evaluation. Given the three key problems， the teaching strategy of “two combinations， one connection”is adopted， and “combining mass entrepreneurship and innovation initiative with practice” is carried out throughout the classroom teaching. The teaching is continuously improved through multiple evaluations so that students can work hard on “truly understand and use”， and jointly create “in-depth courses”. In addition， making full use of scientific research serving to teach， promoting teaching by competition， and promoting the better development of the teaching team are suitable. At the same time， we make full use of learning applications， word clouds， and other digital intelligent means to organically combine online and offline teaching， and strive to build a “learning community” which has achieved remarkable results.关键词：machine vision;teaching reform;teaching mode;solomon learning style;teaching system
- 28
- |
- 9
- |
- 0
发布时间：2022-09-21 - 摘要：Take the ability training of solving complex engineering problems in the background of new engineering for education certification as research objectives， it is proposed that teaching reform method for improving the ability to solve complex engineering problems based on case simulation to solve the issues existing in current comprehensive practice of information security course which is chosen as an example. At the same time， agile scrum is combined into the comprehensive practice teaching. Corresponding training mechanism is designed to strengthen the training of core will quality which is necessary for complex engineering problems solving. From the teaching results and student feedback， it shows that the method is effective which can provide reference for complex engineering problem solving ability training.关键词：complex engineering problems;comprehensive practice;Scrum;core accomplishment
- 26
- |
- 5
- |
- 0
发布时间：2022-09-21 - 摘要：In order to solve the current situation that the training of computer professionals in colleges and universities does not match the industrial demand， it is proposed to link the computer technology and software professional technical qualification （level） examination （software examination） with the training of computer professionals in colleges and universities and the industrial demand. Through the soft examination system qualification setting， evaluation methods， evaluation standards， and the post ability centered assessment knowledge system， colleges and universities are guided to build a complete computer professional training system and curriculum system. The practice results show that integrating the soft test qualification into the course teaching can stimulate the enthusiasm of students to obtain professional titles and professional qualification certificates， enhance the core competitiveness and information processing ability of students， improve the quality of computer talents training in colleges and universities， and provide new methods for computer talents training in colleges and universities.关键词：software examination;computer talent training;substitute examination for title evaluation;professional technical qualification;industrial demand
- 30
- |
- 10
- |
- 0
发布时间：2022-09-21 -
Exploration on the Construction of Interactive Real-time 3D Course for Multi Professional Background AI导读
摘要：With the rapid iteration of digital technology， in order to popularize science and practice the latest technology in the field of digital media， and enable students to apply the new generation of human-computer interaction in their respective fields， a new course of "interactive 3D design， animation and Games" has been opened for students of all majors in the school. From the perspective of teaching practice， the wide coverage of course content， the complex professional background of students who choose courses and different personal abilities lead to the problems of low matching degree of teaching objectives and majors and low completion degree of comprehensive experiments.In view of the problems existing in the course， the teaching mode of different types and stages is redesigned， and the teaching and experimental contents are adjusted. According to the professional direction， the teaching objectives are divided into three types： art， product and control， so that students can selectively learn relevant knowledge according to their respective majors and interests； According to the universality and professionalism of the teaching content， the semester is divided into three stages： primary， intermediate and advanced， to help students from simple to deep， personalized completion of project practice； In terms of teaching methods， we should make full use of online resources to carry out project driven flipped classes， and realize comprehensive and complete projects through group experiments. It provides students with interdisciplinary cooperation experience and cultivates students' design level， autonomous learning and teamwork ability.关键词：interactive3D;real-time rendering;unreal engine;multi professional background;project driven flipped classroom- 10
- |
- 4
- |
- 0
发布时间：2022-09-21 - 摘要：In order to grasp the research progress and dynamic frontiers in the field of online education accurately， so as to provide inspiration for the development of online education in China， with the help of Bicomb and CiteSpace software，make statistical analysis and visual display of the papers in the field of online education included in the core journals of CNKI from 2017 to 2021，outline the research strength， hot topics and the research fronts in this field from the aspects of annual publications， authors， institutions， the statistics of keyword frequency and keyword clustering. The analysis found that the number of published papers in the past five years has generally shown a trend of first decreasing， then increasing and then decreasing，additionally it also showed a lack of cooperation between authors and institutions. The research content mainly focuses on four aspects： teaching reform， learner’s learning process and learning effect， integration with lifelong learning system， and exploration of innovation path. In the future， the research of online education in China may focus on emerging technologies， keep up with cutting-edge topics， change research paradigms， innovate research methods， encourage cross-institutional cooperation and promote research diversification.关键词：online education;knowledge graph;CiteSpace;visual analysi
- 29
- |
- 11
- |
- 0
发布时间：2022-09-21 - 摘要：At present， the so-called Jim Gray's new Moore's law is becoming more and more popular in Chinese academic circles. This study found that most Chinese literature on this law failed to indicate the source of relevant references， nor did it be verified with statistical data. Secondly， this study has traced back to the source， analyzed various Chinese and English literature expressions， and gradually clarified various reasons for the new Moore's law in China， and found no direct evidence of Jim Gray's new Moore's law. Combined with logical analysis and relevant statistics on the growth of global data volume， the expression of the so-called Jim Gregory new Moore's law itself also has obvious loopholes. Therefore， this study speculates that Jim gray did not put forward the so-called Jim gray new Moore's law， which should be a misunderstanding of some views of Jim gray by some domestic scholars. Finally， this study holds that the new Moore's law of the exponential growth of the total global data can be expressed as： the total global data will double every three years.关键词：big data;amount of data;new Moore's law;Jim Gray
- 11
- |
- 3
- |
- 0
发布时间：2022-09-21 -
Research on the Smart City Construction with Fine Management Orientation——A Case Study of Shanghai AI导读
摘要：To realize the fine management of the city is one of the important goals of smart city construction. Therefore， taking Shanghai as an example， this paper analyzes the coping strategies， existing problems and the significance of smart city construction to promote urban fine management. Firstly， it introduces the development status of smart city technology in urban management； Secondly， according to the investigation report of the grass-roots personnel of urban management， identify the problems existing in the technology， management and residents of urban fine management； Finally， drawing on the cases of smart city construction in developed countries， this paper puts forward suggestions on the problems to improve the smart city theory， and provides references for the application scenario design of related technologies. This study summarizes the problems and solutions of smart cities at this stage， in order to provide reference for urban management departments.关键词：smart city;urban fine management;urban construction- 11
- |
- 5
- |
- 0
发布时间：2022-09-21 - 摘要：As China pays increasingly attention to information security， cryptography， as an important branch of information security， its undergraduate education has gradually become a hot spot. It is of great significance to show its development context and research hotspots more intuitively for the better development and research of the discipline. Taking 472 papers in cryptography education and related fields included in CNKI database from 1996 to 2020 as samples， using literature measurement tools CiteSpace and SATI， analyze the time， author， organization and keywords of the literature， and draws the relevant knowledge map and data list. The results show that the research in the field of cryptography education in China has shown an upward trend in recent 25 years. The hot research contents mainly focus on cryptography， information security， information security education， network security and teaching reform.The evolution of cryptography education research can be roughly divided into the initial stage of cryptography education in the early development of cryptography， the large-scale development stage of cryptography education in the transformation of cryptography applications， and the phase of cryptography education in the new era. In the future research of cryptography education， it is necessary to focus on the curriculum reform and strengthen cooperative research.关键词：undergraduate cryptography education;CiteSpace visualization software;SATI visual analysis
- 19
- |
- 6
- |
- 0
发布时间：2022-09-21
- Technical support is provided by Beijing Founder electronics co., LTD 鄂ICP备12016542号-70
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0