最新刊期
卷 21 ， 期 2 ， 2022
- 摘要：Existing recommendation methods based on review text obtain review embeddings through static word vector technologies, and the ambiguity of words leads to the problem of deviations in semantic understanding. Besides, feature splicing strategies cannot balance the impact of user feature and item feature on recommendations. To solve the above problems, propose a review text based adaptive feature extraction recommendation model. First, the model uses the dynamic word embedding technology-BERT as its pretraining model. Next, the model combines Bi-GRU with the attention mechanism for bidirectional feature extraction to fully extract reviews features. Finally, an adaptive feature splicing mechanism is proposed to balance the contribution of user features and item features during interactions. The experimental results show that the minimum mean square error value of the model on the 6 Amazon data sets is 0.678, which is an average increase of 2.42% compared to the performance of the best baseline. The model can not only reduce the impact of word ambiguity in the review text on the recommendation, but also the adaptive feature splicing mechanism can balance the importance of users and products, and improve the accuracy of prediction rating.关键词：recommender system;deep learning;adaptive feature extraction;review text;attention mechanism
- 211
- |
- 56
- |
- 0
发布时间：2022-03-16 - 摘要：In order to solve the problem of production delay caused by uncertain factors in the product manufacturing process, a product cycle prediction method based on improved deep neural network is proposed considering the high data dimension and wide feature space distribution. The weight and bias of the neural network are pre trained by the trestle self encoder and the feature extraction process is completed. The pre training parameters are initialized to the deep neural network for production and manufacturing prediction. In order to avoid the model fitting, L2 regularization and dropout are added. Taking the manufacturing data of a manufacturing industry in a certain quarter as an example, the experimental results verify the effectiveness of the prediction model. Compared with other prediction models, the trestle self coding depth neural network has higher prediction accuracy.关键词：trestle self coding-deep neural network;uncertain factors;manufacturing cycle;prediction model
- 44
- |
- 16
- |
- 0
发布时间：2022-03-16 - 摘要：In order to improve the performance of existing Chinese comment sentiment classification method based on deep learning network, and improve the accuracy and efficiency of Chinese comment sentiment classification. Based on the Tensorflow framework, improved the traditional BERT model, and used the Weibo comment datasets published by Nlpcc2013 and Nlpcc2014 as experimental data. Extracted 3 949 and 14 163 comments with obvious emotional polarity respectively from the two datasets. Input the data into the improved RoBERTa model, extracted the semantics and uses thesigmoid function to classify the positive and negative emotions. Experiment results show that the algorithm accuracy rate, precision rate, recall rate and F1 value of this model are higher, reaching 87.85%, 88.36% and 89.67% on the Nlpcc2013 data set respectively, which is better than the results of the traditional BERT model on various data. The increasement are 1.52%, 0.47%, 2.58% and 1.52%; the performance of various indicators on the Nlpcc2014 data set are increased by 1.30%, 0.54%, 2.32%, and 1.44% respectively compared with the BERT model. This model performs well during the processing of sentiment classification of Chinese comments, and has better outcomes than previous deep learning network models.关键词：deep learning;BERT;RoBERTa;Chinese sentiment classification
- 52
- |
- 24
- |
- 0
发布时间：2022-03-16 - 摘要：Hmong language is one of the main languages of the inhabitants since of the Southwestern era, its phonetic research has problems such as lack of language and text, incomplete phonetic corpus, and regional differences. Therefore, the speech recognition of Hmong language challenges more. Aiming at the problem of imperfect speech corpus, this paper uses pinyin to label the Hmong language, and constructs a corpus of Hmong language voices in different regions; based on the idea of pinyin media transmission, a recognition model of Hmong language isolated words is designed, which uses convolutional neural networks. The network trains and tests the extracted speech features, and solves the problem of Hmong language isolated words speech recognition in the absence of text. In order to verify the effectiveness and stability of the proposed model, the self-built Hmong language speech corpus is used as the experimental data set, the experimental verification is carried out in the same region, different regions and combined regions respectively. The experimental results show that the proposed model is effective for the speech recognition of isolated words in Hmonglanguage, and the effect of speech recognition in the same area is better than that in different areas, The effect of speech experiment in combination area is good.关键词：Hmong language isolated words;Hmong language phonetic tagging;Hmong language voice corpus;speech recognition;convolutional neural network
- 36
- |
- 5
- |
- 0
发布时间：2022-03-16 - 摘要：Research the optimization method based on long and short-term memory neural network model and its application in the prediction of derailment coefficient. Establish train-track simulation scenarios through SIMPACK to obtain data sets required for network training, build a prediction model of derailment coefficients based on long and short-term memory neural networks, use dynamic learning rate and dropout methods to optimize learning rate and network structure, and use the optimized The long and short-term memory neural network predicts the derailment coefficient. The prediction results of the derailment coefficient show that the prediction error of the optimized long-and short-term memory neural network prediction model on the test set is reduced by 23.9% compared with the model before optimization. The dynamic learning rate and dropout method can efficiently optimize the prediction model. The model accurately predicts the change trend of the derailment coefficient, which can provide support for further research.关键词：derailment coefficient prediction;deep learning;dynamic learning rate;Dropout
- 14
- |
- 11
- |
- 0
发布时间：2022-03-16 - 摘要：In order to improve the prediction accuracy of railway freight volume, based on the linear and nonlinear characteristics of freight volume, the freight volume of China from 1980 to 2020 was selected as the research object, and the combined model of ARIMA and BP neural network was proposed. The railway freight volume was modeled and analyzed based on two different combined methods of simple weighting and residual optimization. Compared with the single model of ARIMA and BP neural network, the experimental results show that the combined model of ARIMA and BP residual optimization has higher accuracy, and the evaluation indexes MAE, MAPE and RMSE are 5 135.90, 0.020 9 and 8 462.49, respectively. The combined model based on ARIMA-BP residual optimization has realized a good prediction of freight volume, which is of great significance for the railway transportation sector to make relevant policies flexibly.关键词：ARIMA;freight volume forecast;portfolio model;simple weighted;residual optimization
- 16
- |
- 4
- |
- 0
发布时间：2022-03-16 - 摘要：Aiming at the problems of poor polysemous word parsing effect and poor generalization ability in Chinese location semantic analysis using current neural network models, propose a Chinese location semantic analysis method based on the BERT-BiLSTM-CRF model. This method first uses the BERT pre-training model to pre-train the Chinese location information to obtain the context information in all layers, enhance the semantic representation ability of the Chinese location information, and then extract the feature information of the vector through the BiLSTM model, and finally decode it through the CRF model to obtain the global Optimal labeling sequence. Experimental results show that on the basis of different numbers and regions of Chinese location information data sets, the word segmentation accuracy and F1 value of the BERT-BiLSTM-CRF model on all test sets are better than the current commonly used neural network models, and the highest F1 achieved are 93.91% and 93.96%, respectively. Using the BERT-BiLSTM-CRF model to perform semantic analysis on Chinese location information not only effectively improves the accuracy of Chinese location segmentation and polysemous word analysis, but also has better generalization capabilities.关键词：semantic analysis of Chinese location;BERT pre-training model;BiLSTM model;CRF model
- 25
- |
- 9
- |
- 0
发布时间：2022-03-16 - 摘要：Sentiment analyses of online reviews for travelling are of great significance for understanding tourists’preferences. However, the existing emotional analysis models of tourism evaluation rarely highlight the core content of tourists’ emotional description. In order to analyze emotional tendency of tourists’evaluation more accurately, proposed a sentiment analysis model for travelling, which consists of the part-of-speech attention mechanism and Bi-directional Long Short-Term Memory (Bi-LSTM) neural network, where the model adopts Bi-LSTM to capture contextual semantic information, and makes full use of the relationship between part-of-speech of words and emotional polarity words in a sentence, and focus on the core part of the emotional description contents instead of relying on artificially organized emotional domain dic-tionaries. The results show that the sentiment analysis method proposed has higher accuracy (91.84%), precision (90.63%), recall (92.33%), and F1 score (91.47%) than that of SAM+Bi-LSTM, thus improving by 4.3%, 1.02%, 5.12%, and 3.08%, respectively.关键词：sentiment analysis;Bi-LSTM;attention mechanism;artificial intelligence
- 20
- |
- 8
- |
- 0
发布时间：2022-03-16 - 摘要：Due to the expansion of the traditional rapid improved algorithm of random tree (RRT*) in the treatment of the autonomous underwater robot AUV (Autonomous Underwater Vehicle) three-dimensional path planning problem of the existed algorithms of slow convergence speed, rough planning path, propose a modified NT-RRT*algorithm (normal sampling, triangle cut the rapid expansion of random tree algorithm). In order to improve the convergence speed of the algorithm, a spatial sampling strategy based on normal distribution is proposed to replace the global uniform random sampling in RRT*algorithm. In addition, a geometric pruning algorithm based on triangle inequality is introduced. It reduces the length of nodes and paths in the process of random tree expansion. Finally, the simulation is carried out in a three-dimensional underwater environment with terrain obstacles and floating obstacles. The simulation results show that the number of nodes in random expansion of the improved algorithm is reduced to 15%, the algorithm planning time is shortened to 20% of the original, the planned route length is about half of the original, and the path is smooth and the curvature is small, which shows that the improved NT-RRT*algorithm is obviously better than the RRT*algorithm in terms of convergence speed and path length.关键词：autonomous underwater vehicle;RRT*algorithm;normal distribution sampling strategy;geometric pruning algorithm based on triangle inequality
- 19
- |
- 3
- |
- 0
发布时间：2022-03-16 - 摘要：The model structure of traditional speech emotion recognition algorithms was usually relatively simple, and it required a sufficient amount of training data to had certain practicability. However, in practical applications, the data set available for training was often only dozens of sentences. In order to alleviate this contradiction, a speech emotion depth transfer recognition algorithm is proposed. The source domain data set was determined through operations such as attribute analysis table construction, feature extraction, and correlation calculation, and on this basis, pre-training model construction and transfer learning training were performed to form a speech emotion classification model. Experimental results show that this algorithm can solve the problem of low-resource data sets easily falling into local optimality, the recognition rate is more than 60%, which has certain advantages compared with BLSTM and CNN models.关键词：transfer learning;speech emotion;convolutional neural network;deep learning;pattern recognition
- 15
- |
- 5
- |
- 0
发布时间：2022-03-16 - 摘要：The traditional single-domain recommendation algorithm has problems of data sparsity and cold start, and it makes recommendation based on group classification in the same domain, so it cannot really achieve personalized and accurate recommendation. In the network environment, learning resources present a variety of forms, and the rating data of learners in each field is more sparse. In view of this, we proposed a cross-domain learning resource recommendation algorithm of migrating rating pattern, which first obtains learners'rating pattern andresources’rated pattern from learners’rating data of auxiliary domains based on the idea of codebook clustering, and then transfer them to the target domain in order to fill in the deficiency of the rating data in the target field. Finally, this algorithm integrates the learners, learning resources in the target field, rating pattern and rated pattern by factorization machines (FMs) to make accurate and personalized learning resources recommendation for learners. The experimental results show that this method can solve the problems of data sparsity and cold start in single domain learning resource recommendation, and improve the accuracy of personalized learning resource recommendation.关键词：transfer learning;codebook-clustering;factorization machines;rating mode;cross-domain recommendation
- 15
- |
- 7
- |
- 0
发布时间：2022-03-16 - 摘要：With the development of the Internet, the evaluation of online goods and services has a great impact on consumers' decision-making. Therefore, a false comment detection method based on CNN-LSTM model and transfer learning is proposed. The model uses CNN-LSTM algorithm on the data set, combines Doc2vec and TF-IDF method to quantify the evaluation text as features, and finally uses Glove data set for transfer learning. The experimental results show that the proposed detection model has higher detection and recognition accuracy than other false comment detection models, and the accuracy rate reaches 93.1% in the experimental test of the English Hotel comment data set of deceptive-opinion-spam-corpus.关键词：text analysis;CNN;LSTM;false comment recognition
- 25
- |
- 8
- |
- 0
发布时间：2022-03-16 - 摘要：Aiming at solving the problem of tag collision in radio frequency identification (RFID) system, proposed an adaptive query tree anti-collision algorithm based on mapping sequence code (MAQT). The algorithm is based on the characteristics of the three consecutive bits headed by the highest collision bit, and combining with the unique mapping relationship, the algorithm can accurately infer the query prefix existing in the current query slot, so as to eliminate the idle slots and reduce the collision slots. Simulation results show that, When the number of tags to be identified reaches 1 000, MAQT algorithm can reduce the total slot number by 10.8% compared with GBAQT algorithm, and the throughput can be maintained at about 0.61. The algorithm can reduce the total number of time slots and improve the system efficiency.关键词：radio frequency identification (RFID);anti-collision algorithm;query tree;mapping relationship
- 21
- |
- 7
- |
- 0
发布时间：2022-03-16 - 摘要：In order to conduct a comprehensive visual analysis of online social network, the method of persistent homology was used to calculate and analyze the communication of online social network. The online social network is regarded as a distance subspace and different subcomplex of this space are constructed. Then the corresponding Betti number and persistence graph are obtained by calculating the homology of simple complex, and the topological characteristics of social network are obtained based on Betti number and persistence diagram. Finally, the communication characteristics of several different types of online social networks are analyzed, and persistent homology is applied to online social network communication. In this way, the persistent homology information of online social networks is obtained, and a new method to describe the communication of online social networks is provided.关键词：online social network;persistent homology;Betti number;persistence diagram;spread
- 20
- |
- 5
- |
- 0
发布时间：2022-03-16 - 摘要：The on-line estimation accuracy of battery capacity affects the accuracy of vehicle driving range and power output. Considering the complex practical on-board environment, improving the accuracy of online capacity estimation is still a challenge. The traditional incremental SOC (stage of charge) -capacity method usually uses two different SOC points as the incremental SOC. The errors of the selected SOC points inevitably lead to inaccurate capacity estimation. A capacity estimation method based on OCV (open circuit voltage) identification and IRLS (iteratively reweighted least squares) algorithm method is proposed. Firstly, based on the first-order RC model, the forgetting factor recursive least squares algorithm is used to identify the OCV online. Then, the full interval SOC estimation result can be obtained indirectly. Finally, based on the idea of the incremental SOC-capacity method, the capacity is estimated using the full interval SOC estimation with IRLS algorithm. Through the designed aging experiment, the capacity estimation accuracy is verified under different life conditions. The results show that the estimation accuracy of this method is high and the error does not exceed 2%. This method avoids the randomness of the SOC point selection, and thus solves the problem of possible outliers in the SOC estimation or the unknown variance of the random error of the SOC gain.关键词：battery life;capacity estimation;open circuit voltage;least squares method;online application
- 13
- |
- 3
- |
- 0
发布时间：2022-03-16 - 摘要：With the increasing demand for liquid crystal display, the efficiency of display production has become even more important. At present, there are some problems in the display automatic production line, such as low automation level and large labor cost. Therefore, increasing the mechanical automation level of the display assembly line can improve the production efficiency of the display, at the same time, it can also reduce the input of labor costs. Display automated assembly line, the author of this paper designed a kind of linkage of the positioning, design principle of the mechanism for mathematical modeling, complete the positioning of the kinematics analysis, and through the post-processing module of virtual simulation software of slider velocity and acceleration curve, the curve smooth transition, no sharp point, It shows that the designed centering and positioning mechanism has no great impact on the LCD during movement, which can be used as the theoretical basis for the automatic production line of display. Finally, through the inverse kinematics analysis of the alignment positioning mechanism, the rationality of the designed mechanism is verified by the simulation test of virtual prototype.关键词：liquid crystal display;center positioning mechanism;kinematic analysis;virtual prototype simulation
- 12
- |
- 4
- |
- 0
发布时间：2022-03-16 - 摘要：By analyzing the stress shunt of the femoral-screw-plate system, provide guidance for determining a reasonable screw fixation position during the treatment of femoral shaft fractures. Import the DICOM format image data of the femur obtained by the CT scan into the Mimics software, and obtain the geometric model of the complete femur through three-dimensional reconstruction; establish the three-dimensional model of the bone plate and the screw in the SolidWorks software, and combine the femur and the bone plate with three different layouts. The screws were assembled and imported into HyperWorks software to simulate the mid-femoral shaft transverse fracture to analyze the stress shunt of the femoral-screw-plate system under the condition of standing with two legs. The stress of the femur is concentrated at the screw connection, and the stress of the bone plate is concentrated at the fracture line; When the six screws are fixed on the eight-hole bone plate and the two middle holes are empty, the working length is longer and the stress of the bone plate is the smallest. The stress ratio of the bone plate at the fracture is 87.46%. The stress shunt study of the femoral-screw-plate system is helpful to determine the reasonable screw layout, and the results obtained are consistent with the clinical treatment of femoral shaft fracture fixation. It can provide theoretical basis and guidance for further research on the biomechanical properties and clinical treatment of fracture internal fixation.关键词：internal fixation system;finite element analysis;stress shunt
- 10
- |
- 3
- |
- 0
发布时间：2022-03-16 - 摘要：Scheduling problem is an important combinatorial optimization problem, which is widely used in production planning, computer control and other fields. Research on the scheduling problem of three parallel machines with service level, each machine and each job has a level label, each job can only be processed by a machine whose service level is not higher than that of the job. The goal is to minimize the maximum machine completion time. Use a new algorithm to solve the load balancing problem of three machines under the offline state level constraint. A 4/3 approximation algorithm is proposed for three parallel machines with level constraints of 1, 2 and 2; For three parallel machines with level constraints of 1, 1 and 2, a 2 approximation algorithm is proposed.关键词：hierarchical constraints;scheduling problem;approximate ratio;objective function
- 9
- |
- 3
- |
- 0
发布时间：2022-03-16 - 摘要：The injury caused by falls threatens the life safety of the elderly. A real-time detection method using millimeter wave radar is proposed for real-time monitoring of falls that cause serious damage to human body. The sequence of position coordinates is first mean filtered and then the approximate motion pattern of the monitored object is obtained by segmented linear fitting. The average velocity of the target in the height direction and horizontal direction, as well as the change in horizontal distance, are calculated and compared with the corresponding thresholds to complete the detection and classification. The experimental results show that the accuracy of fall detection is 95.4%, the recognition accuracy of falls during walking was 100%; and the recognition rates of falling while squatting and falling when getting up were 85.71% and 80% respectively. Compared with the method of setting only the height change rate as the threshold, setting multiple thresholds improves the accuracy of fall detection and fall mode recognition by 14.6% and 23.5% respectively.关键词：millimeter wave radar;signal processing;fall detection;fall mode recognition;classification algorithm
- 11
- |
- 4
- |
- 0
发布时间：2022-03-16 - 摘要：The default resource scheduling method of the Kubernetes cluster was static, as the working nodes in Kubernetes continued to run, the problem of cluster load imbalance would occur. In response to the above problems, a dynamic scheduling mechanism for Kubernetes clusters based on Pod migration was designed. The mechanism first collected server and Pod application resource usage through the monitoring module, and uses AHP to evaluate its load; then through the definition of a sliding window, high-load working nodes were filtered to form a set of Pods to be migrated; finally selected the target working node for the Pod to be migrated base on The low-load node principle and adopted CRIU technology to realize the Pod migration. Experimental results show that compared with priority resource scheduling, the dynamic resource scheduling mechanism has an average increase of 17% in load balance and an average reduction of 88.3% in the number of pod migrations.关键词：Kubernetes;Pod migration;dynamic resource scheduling;AHP;load balancing;cloud computing
- 14
- |
- 8
- |
- 0
发布时间：2022-03-16 - 摘要：With the deepening of the information construction of university education, the all-around intelligent and digitalized management of university campus laboratory is also imperative. In view of the characteristics of various kinds of experimental equipment, large number of students, complex experimental documents and data, and special requirements of teaching management existing in professional practice teaching in colleges and universities, a set of cloud-side collaborative server architecture was designed around the Web-side control and management system of college laboratories, combining the advantages of cloud and edge processing. This paper mainly analyzes the high-performance asynchronous communication mode between the experimental equipment and the edge server, edge server and cloud server, in order to realize real-time data interaction and real-time control of the experimental platform by the Web side. WeChat small program side laboratory management system has been developed to enhance the intelligence of the laboratory and promote the realization of the intelligent campus.关键词：Internet of Things;Netty;cloud edge collaboration;smart campus
- 26
- |
- 31
- |
- 0
发布时间：2022-03-16 - 摘要：In order to solve the problem that the existing bearing fault diagnosis methods based on deep learning are not suitable for small sample data, too many super parameters and high computational cost, a bearing fault diagnosis model based on Deep SGD-Forest (DSGDF) is proposed. The method is based on the deep forest model, and the gradient descent algorithm is integrated into the multi-grained scanning and the cascade layer to improve the convergence speed of the model. Experimental results on the bearing dataset of Case Western Reserve University show that the average diagnostic accuracy of the DSGDF model is more than 99%. Compared with the deep forest model, the DSGDF model has faster convergence speed and less computational overhead compared with other deep learning models based on neural networks.关键词：deep learning;DSGDF;bearing;fault diagnosis
- 14
- |
- 8
- |
- 0
发布时间：2022-03-16 - 摘要：To explore the thermal insulation performance of composite concrete self-insulating blocks, COMSOL Multiphysics software was used to simulate the heat and moisture transfer process of blocks under different boundary conditions and initial conditions, the temperature distribution inside the block was obtained. Simulation results showed that, under steady-state boundary conditions, when the internal relative humidity of the block is lower than 50%, the influence on the internal heat transfer rate of the block is not obvious. When the relative humidity is higher than 50%, the internal heat transfer rate of the block is significantly affected; Under unstable boundary conditions, there is a certain time lag in the heat and moisture transfer inside the block. The more water contained in the block, the better the thermal conductivity of the block has. If the thermal insulation performance of the block is improved, not only the block should be treated with moisture insulation, but also the relative humidity of the external environment should be reduced to prevent the block from absorbing moisture and increasing its internal moisture.关键词：coupled heat and moisture;simulation;composite block
- 12
- |
- 3
- |
- 0
发布时间：2022-03-16 - 摘要：Due to different levels of customs inspectors in wood classification, some rare and endangered woods are often forged as common woods for avoiding national supervision. To provide a convenient automatic identification tool of wood species for customs inspectors, a remote identification system of wood images based on Web was developed. The wood image remote identification system includes Web side, wood database and server side, where the server side includes application server and wood image recognition server. The algorithm model is deployed on the wood image server, the identification server is requested through the application server, and the identification server calls the algorithm model identification results. When a customs inspector encounters a wood he does not recognize, he uploads the wood image and the backend server gives feedback and displays the information after the recognition is completed. The system provides the wood information search, the intelligent wood species identification and expert online support for 120 species of common woods in import and export trade, 120 types of wood images have an average recognition accuracy rate of 95.7%, which provides a support tool for identifying quickly wood species to customs inspectors and greatly saves economic and time costs.关键词：Web;wood images;remote identification;deep learning;image identification
- 33
- |
- 72
- |
- 0
发布时间：2022-03-16 - 摘要：When air quality monitoring is performed in an indoor home environment, there are often multiple rooms that need to collect air quality data at the same time. Therefore, design a distributed multi-node indoor air quality monitoring system (AQMS) which uses the Message Queuing Telemetry Transport (MQTT) protocol. Transplanted with RT-Thread real-time operating system, the monitoring node is composed of DS18B20 temperature sensor, MQ-7 CO sensor and GP2Y1014AU microparticulate sensor. Multi-node monitoring data is gathered to the gateway node through the nRF24L01 data transmission module. The data then is uploaded to the OneNET IoT cloud platform by the WiFi data transmission module ESP8266 using the MQTT protocol. The cloud platform can view the air quality data uploaded by each monitoring node in real time, and can also control the 4 relays connected to the gateway node on the cloud platform to realize remote operation of household appliances and air quality alarm information can be sent to the user's mobile phone. The experimental results show that the indoor AQMS designed operates reliably.关键词：indoor air quality monitor;IOT;MQTT protocol;RT-Thread OS;OneNET cloud platform
- 13
- |
- 3
- |
- 0
发布时间：2022-03-16 -
Design of Remote Experimental Platform for Hardware Courses Based on MOOC and FPGA Technologies AI导读
摘要：In traditional courses of computer hardware, the experimental teaching is limited by time, space and hardware equipment. To solve the problem, it is necessary to explore a new teaching equipment based on online interactive leaning, and design and develop a remote FPGA virtual simulation experiment platform based on MOOC for teaching, which supports online learning and experiments anytime and anywhere, solves the problem of time and space constraints, and improves the utilization of experimental equipment. It supports the online experiment of several hardware courses taught by FPGA, and realizes the connection between courses. In addition, the statistical model provides experimental data analysis to help teachers analyze their students’ability for adjusting experimental content. During the COVID-19 epidemic, the platform supported experimental courses such as digital circuit and computer composition principle. The results show that the platform provides students with sufficient opportunities for hardware experiments to cultivate and improve the design ability of bottom system.关键词：experimental teaching platform;FPGA online experiment;MOOC platform;hardware courses;systematic ability training- 21
- |
- 6
- |
- 0
发布时间：2022-03-16 - 摘要：In order to solve the problem of intelligent garbage collection and monitoring, an intelligent garbage monitoring system equipment was designed based on WeChat applet. The system is consists of an intelligent garbage data acquisition equipment, a WeChat Applet and a Cloud server database. The cloud server database is used for data storage and transfer, through which the data obtained from the monitoring system will be sent to the WeChat applet. The system’s main functions are speech recognition, overflow detection, fire detection, alarm function and so on. Then with the help of the WeChat applet, cleaners can check the real-time status of garbage cans, and maintenance of garbage equipment and garbage collection timely, thus realize the effective management of garbage equipment and efficient recycling of garbage.关键词：intelligent garbage can;Raspberry Pi;cloud server;WeChat applet
- 18
- |
- 6
- |
- 0
发布时间：2022-03-16 - 摘要：In the structural design of modern mechanical products, high precision, light weight and stability have become the key design requirements. Manipulator is a key component of construction machinery facilities, improving the dynamic characteristics of the manipulator and meeting the individual needs can effectively improve the quality, efficiency and safety of construction machinery. Therefore, the 6-DOF manipulator is taken as the research object, SolidWorks software is used to model the 3D entity of the robot arm, and C#programming language is used to carry out the secondary development of SolidWorks, design the visual operation interface, and create the parametric design system of the manipulator. The system can dynamically change the size of the model parts according to the parameters calculated by the structure design, and realize the automatic assembly of the manipulator model. Under the premise of ensuring the quality of product development, the system greatly improves the design efficiency and saves the development cost.关键词：manipulator;secondary development;parametric design;three-dimensionalsolid model building
- 9
- |
- 3
- |
- 0
发布时间：2022-03-16 - 摘要：In the era of cloud computing and big data, digital signature technology plays an irreplaceable role in cryptography. Digital signatures can not only guarantee the authenticity and non-repudiation of data, but are now more used to implement public audits of the integrity of various outsourced data in cloud storage. Starting from the actual needs of cloud storage, this article designs a certificateless linear homomorphic aggregation signature scheme suitable for public batch verification of multi-user data in cloud storage. The security of the signature scheme can resist internal signatories from the aggregation. Collusion attack"; in terms of efficiency, this scheme can compress and aggregate the signatures generated by various data of different users. Through simulation experiments under the JPBC library, it shows that this scheme improves the verification speed nearly 10 times compared with ordinary certificateless digital signatures. Besides, in terms of storage cost, the scheme only needs to store a digital signature after homomorphic aggregation of all data. Compared with the traditional"one piece of data, one signature"storage method, it fundamentally saves users's torage space on the cloud.关键词：information safety;cryptography;linear homomorphic signature;aggregate signature;certificateless cryptosystem
- 16
- |
- 4
- |
- 0
发布时间：2022-03-16 - 摘要：Differential privacy is a privacy protection model that can provide rigorous mathematical proof, and in view of the poor application of traditional differential privacy protection methods in mixed data sets and the problem of data availability after processing, propose a differential privacy protection algorithm for adaptive clustering of mixed data sets. By combining the characteristics of fast clustering and k-prototype clustering algorithm, the distance calculation of the properties of different data types is realized by different degree calculation methods for mixed data sets, and then the Laplace mechanism and exponential mechanism are used in the cluster center after clustering, and finally the differential privacy protection of mixed data sets is realized by combining the combined nature of differential privacy. The experimental results show that the data error rate is reduced by 15% compared with the same privacy budget compared with the traditional method of processing mixed data sets, which proves the availability of processing results. Under the fixed privacy budget, changing the initial clustering number, the privacy protection degree of this method can reach 96%, which proves the effectiveness of this algorithm in processing mixed data sets.关键词：differential privacy;hybrid data set;clustering algorithms;privacy protection;self-adaptation clustering
- 13
- |
- 6
- |
- 0
发布时间：2022-03-16 -
Classification of Hyperspectral Images with Feature Perception and Collaborative Representation AI导读
摘要：In order to solve the problems of noise, complex spatial structure and complex spectral information in hyperspectral images, and improve the ability of noise processing and spatial recognition of the classification algorithm, a hyperspectral image classification method based on feature perception and collaborative representation was proposed. Firstly, the algorithm uses adaptive weighting to reconstruct the image. Then, the spatial features are perceived by calculating the spatial bias matrix, and the spectral features are perceived by calculating the spectral bias matrix. Finally, according to the principle of minimum error, the category information of the test samples is determined. The experimental results on the standard data sets Pavia University and Salinas show that the overall accuracy on the two data sets reaches 98.96% and 99.63%. Compared with the advanced classification algorithms, the algorithm has better effects in terms of smoothing noise, perceptual features and spatial recognition, which has better classification performance.关键词：collaborative representation;hyperspectral image;classification;feature perception;regularization- 14
- |
- 6
- |
- 0
发布时间：2022-03-16 - 摘要：Aiming at the problems of traditional dark channel priori algorithm after image defogging, such as incomplete defogging, halo effect at the edge of the scene and slow processing speed, an image defogging algorithm based on bottom hat transformation and image fusion is proposed. Firstly, down-sampling the fog map to speed up the algorithm processing speed; secondly, according to the characteristics of the bottom hat transformation to obtain a better contour map, first perform the bottom hat transformation on the minimum channel image, and then merge with the gray-scale inversion image, thereby optimize the edge of the dark channel image; thirdly, merge the transmittance after bilateral filtering and opening-closing operations with the transmittance obtained from the improved dark channel to achieve an accurate estimate of the transmittance, and then use the guided filter to refine the transmittance; finally, use the atmospheric scattering model to obtain the restored image. The experimental results show that the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) of the image processed by this algorithm are increased by 4.3% and 8.5%, respectively, while the processing time is reduced by 98.5%.关键词：dark channel prior;down-sampling;bottom hat transform;image fusion;bilateral filtering
- 16
- |
- 6
- |
- 0
发布时间：2022-03-16 - 摘要：In order to improve the detection efficiency of cigarette defects in the process of cigarette production and to find the reasons of cigarette defects, propose a method for classification and location of cigarette defects based on multi-neural networks. This method first performs image preprocessing on the cigarette data set, then uses the improved VGG19 network to predicting whether there are defects in the cigarettes, using the YOLOV3 network to locate and determine of the cigarettes which are predicted have defects in the cigarette data. Finally, according to the cigarette statistical results of the defect categories inversely infer the causes of the defects in the cigarettes and locate the wrong process in time, so as to realize the detection of cigarette defects. Research shows that the detection method has high detection accuracy, accurate defect category analysis and accurate defect positioning.关键词：cigarette defects;neural network;VGG19;YOLOv3
- 14
- |
- 4
- |
- 0
发布时间：2022-03-16 - 摘要：Based on the relevant literature of educational informatization leadership research in CNKI "Chinese Academic Journal Network Publishing Database" as the data source, this paper uses SATI3.2 and SPSS19.0 for statistical analysis, and uses the combination of quantitative research and qualitative research to analyze the time distribution, author distribution and ownership unit distribution of the literature, and carries out visual analysis based on high-frequency keywords. The results show that the leadership research of educational informatization in China mainly focuses on the following three aspects: "educational informatics 2.0" background principal informatics leadership promotion research, principal informatics leadership’s role of school informatics development research, teacher informatics leadership research. At present, the research on educational informatization leadership is still in the development stage, the future research will develop in the direction of decentralization, with equal emphasis on normative research and empirical research, leadership governance capability and teachers’information teaching leadership in the smart environment will become the research trend in the future.关键词：education informatization;informatization leadership;cluster analysis;social network;current situation analysis;trend research
- 31
- |
- 9
- |
- 0
发布时间：2022-03-16 - 摘要：In view of the shortcomings of high level language programming course teaching under the background of emerging engineering, analyzed the key problems to be solved in high level language programming, combined with the characteristics of programming course, an intelligent teaching framework of high level language programming was constructed, which has been implemented in the teaching process. This method has achieved good results in the informatization of teaching process, the digitalization of teaching management and the intelligentization of teaching decision-making, the innovation and integration of intelligent education has promoted the overall improvement of teaching quality.关键词：emerging engineering;high level language programming;intelligent teaching;teaching framework design
- 24
- |
- 8
- |
- 0
发布时间：2022-03-16 - 摘要：With the rapid development of information, the technology of artificial intelligence has become the mainstream research. How to improve the training of innovative talents has become the responsibility of universities. Based on the analysis the positive role of "Challenge Cup" competition in the cultivation of innovative talents in the era of artificial intelligence, this paper firstly takes Nanjing University of Posts and Telecommunications as the platform, to research and practice the "six in one" innovative talents training mode based on "Challenge Cup". And then, it takes the medical image segmentation based on deep learning as an example to verify the effect of "Challenge Cup" on the cultivation of innovative talents. Finally, it briefly introduces the effect of training mode of innovative talent of the school relying on "Challenge Cup" and other innovative competitions and projects.关键词：six in one;Challenge Cup;artificial intelligence;innovation talent;talent cultivation
- 29
- |
- 6
- |
- 0
发布时间：2022-03-16 - 摘要：In order to improve the shortcomings of inaccurate positioning and lack of resources in the cultivation of innovative talents in information discipline of the existing local comprehensive universities, according to the discipline characteristics and diversified talent training objectives, taking Xiangtan University as an example, a multi-layer training platform at the national, provincial and school levels was built for the training of scientific research and application-oriented compound talents. Based on the tutorial system, the multi-agent collaborative training mode of inter university scientific research cooperation and school-enterprise cooperation was realized through the second class, stepwise and multi-disciplinary cooperation, scientific research, enterprise cooperation and competition as a whole to achieve multi-ways collaborative training. The practice results show that it can fully mobilize students’innovative passion, stimulate students’innovative thinking, promote the combination of scientific research and practice, serve the local economy, and cultivate talents according to the national strategic needs. It is of great significance to innovate the training mode of information talents in local colleges and universities and promote the sustainable development of students.关键词：local comprehensive university;information science;talent training;cultivating mode;collaborative innovation
- 17
- |
- 7
- |
- 0
发布时间：2022-03-16 - 摘要：Aiming at the cultivation of comprehensive quality, firstly analyzes the current situation and existing problems of data visualization course teaching. Then, from the expansion of curriculum content and innovative design of assessment methods, puts forward the curriculum reform ideas and implementation means for the cultivation of comprehensive quality, and introduces the reform practice process. Finally, the effect of curriculum reform is illustrated by the actual situation of personnel training. In the past four years, 254 students have been trained in the course of data visualization, and 67 visualization project cases have been formed. Among the 169 students who have graduated, 82 students continue to study in well-known universities at home and abroad, accounting for 48.6% of the total. The teaching reform program has achieved good results in the cultivation of comprehensive quality of talents and the construction of resourses.关键词：comprehensive quality cultivation;data visualization;curriculum reform;computer education
- 11
- |
- 5
- |
- 0
发布时间：2022-03-16 - 摘要：With the accelerating pace of education modernization, information technology continues to promote changes in the field of education. By selecting three types of forward-looking information technology, discuss VR, ARand MR’s potential, effects challenges and promotion in educational applications. The era of education information needs the common development of people and technology. Therefore, it is still necessary to further improve teachers’information technology application ability and strengthen the cultivation of students’information awareness in order to better realize the educational value of information technology.关键词：education informationization;information technology;virtual reality technology;augmented reality technology;mixed reality technology
- 15
- |
- 8
- |
- 0
发布时间：2022-03-16 -
Research on the Curriculum Ideological and Political Education of Discrete Mathematics Teaching AI导读
摘要：In the implementation of the whole process of education and all-round education, college teachers have the responsibility of ideological and political education, and curriculum ideological and political education plays an important role. Discrete mathematics is a basic course for computer majors. Teachers should dig deep into the ideological and political education resources in discrete mathematics course, run through the content of ideological and political education in the teaching process, and play the main channel function of curriculum education. This paper mainly discusses the construction of ideological and political resources in discrete mathematics teaching, the existing problems in teaching, and puts forward the corresponding teaching cases for the teaching content of discrete mathematics. The teaching process of ideological and political education in discrete mathematics course is designed by the combination of online and offline, which can provide inspiration for the integration of ideological and political education into the teaching of other professional courses.关键词：curriculum ideological and political education;discrete mathematics;teaching cases- 10
- |
- 6
- |
- 0
发布时间：2022-03-16 - 摘要：With the advent of the era of knowledge economy, the importance of deep learning ability in the cultivation of talents in colleges and universities has become increasingly prominent. In the cloud computing course, the A/B group control research method is used for random grouping, and the difference analysis and regression analysis are used to analyze the influence of the online peer questioning strategy on the deep learning ability. Research shows that the peer questioning strategy based on the 5E teaching model in the online environment can effectively improve the deep learning ability of college students; the learner's ability to ask questions has a positive effect on their deep learning ability. This conclusion can be used as the focal point of computer teaching reform, also as to provide reference for the training of computer talents.关键词：deep learning ability;peer questioning ability;peer questioning strategy
- 12
- |
- 6
- |
- 0
发布时间：2022-03-16 - 摘要：In order to solve the problem that the traditional experimental teaching of programming courses lacks efficiency and the instructor cannot verify all the students’computer codes, design and implement the OJ system which integrates online programming, online code evaluation, experimental report generation, ranking statistics and other functions. On this basis, the experimental blended teaching model is proposed, and the design of blended teaching activities is discussed. Through teaching practice, it can effectively improve students’learning enthusiasm and programming ability, and the teaching effect is good.关键词：programming design;experimental teaching;online program;online evaluation;blended teaching
- 13
- |
- 4
- |
- 0
发布时间：2022-03-16 -
Teaching Reform of High-level Language Programming under the Background of Engineering Certification AI导读
摘要：In order to solve some problems existing in the traditional classroom of high-level language programming, carry out curriculum reform from four aspects: auxiliary teaching resources, teaching content, teaching methods and curriculum evaluation system. The innovative teaching methods, such as online and offline mixed teaching, case teaching, graded and stratified teaching, and flipped teaching, are adopted to realize the effective combination of knowledge imparting and value guidance, theory imparting and ability training, and the teaching concept of emphasizing results more than process. According to the result of implementation, the reform effectively improves the teaching effect of high-level language programming and meets the training requirements engineering education certification.关键词：engineering education professional certification;high-level language programming;teaching reform;ideological and political theories teaching in all courses- 15
- |
- 7
- |
- 0
发布时间：2022-03-16 - 摘要：In the first half of 2020, due to the control of the new crown pneumonia, online teaching has shouldered the burden and has become the mainstream teaching method in my country during the epidemic. Under this situation, designed a set of scientific online teaching programs. Compares and predicts the overall teaching effect of the oneline and offline teaching modes by establishing objective and subjective evaluation index systems, using BP neural network as a teaching tool. Results show that, online teaching has relatively obvious improvement in independent learning ability and listening efficiency, which has played a guiding role in the combination of the two teaching modes.关键词：online teaching;offline teaching;evaluation index system;BP neural network
- 13
- |
- 6
- |
- 0
发布时间：2022-03-16 - 摘要：Under the guidance of "all-round education" and "all-process education", colleges and universities actively explore the integration mechanism of professional courses and political courses. Try our best to cultivate students into quality talents with moral cultivation and all-round development of knowledge ability, and realize the unity of "knowledge imparting" and "value leading". Based on the teaching reform of "ideological and political course", this paper studies the current situation of Python course at home and abroad, and analyzes the teachers, students and teaching environment of Python course of political course, attempts to use the action-oriented teaching method to design the teaching material of Python programming fundamentals into a teaching example base of individual, society and state aspects of political education according to the course standard, providing references for teachers to integrate ideological and political ideals into professional courses.关键词：ideological and political theories teaching;action-oriented;Python program design;teaching reform
- 16
- |
- 8
- |
- 0
发布时间：2022-03-16 - 摘要：To solve the problem that the professional theory and reality of computer professional degree graduate students are out of line, the training system should be developed from emphasizing theory to balancing theory and practice; combined with the characteristics of computer specialty, under the guidance of "practical education", the new idea of practice education should be implemented, and the "six elements" achievement oriented enterprise practice system and ability training should be the main line in the whole process. In order to promote the reform and development of education in the new era through practice, give full play to the educational advantages of socialist education with Chinese characteristics.关键词：computer major;graduate education;practice education
- 8
- |
- 5
- |
- 0
发布时间：2022-03-16
- 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