最新刊期
卷 22 ， 期 5 ， 2023
-
Automatic Extraction Model of Multi-document Summarization Based on Improved TextRank Algorithm AI导读
摘要：Multi-document summarization extracts general information from multiple documents of the same topic through natural language processing technology， which can effectively alleviate the current information load problem and help users quickly and accurately obtain the core content of the original text. An improved multi-document extractive summarization model based on the TextRank algorithm according to the characteristics of Chinese text was constructed. Firstly， the pre-training Word2Vec word vector model is fused with SIF method on Chinese Wikipedia corpus to obtain sentence vectors of all sentences in the document. Then， with the cosine similarity， the edge relations between TextRank sentences are created. Finally， the MMR algorithm is used to deal with the redundancy of abstract sentences to obtain a comprehensive and diverse summary. The performance of the model is evaluated by ROUGE-N evaluation index， and the comparative experimental results show that the values of ROUGE-1， ROUGE-2， and ROUGE-L indexes of the proposed model are 0.549， 0.322， and 0.357， which are better than the traditional TextRank method and Word2vec （experimental sample corpus） + TextRank + MMR model， with higher quality.关键词：multi-document summarization;extractive summarization;TextRank algorithm;Word2vec;SIF- 193
- |
- 54
- |
- 0
发布时间：2023-05-29 - 摘要：Aiming at the problem that the traditional click-through rate prediction model is insufficient in extracting user historical behavior features and ignores low-order features， a new click-through rate prediction model TDFA （TransDeepFM-AttentionBased） is proposed. Firstly， a Transformer network is constructed based on the self-attention mechanism to obtain the correlation information among the user's historical behaviors， and then the historical behaviors are associated with the target item to obtain the user's historical behavior features. In order to get higher-order combinatorial features， the historical behavior features， user information， context information and target item information are combined into the multi-layer neural network. Second， the low-level features are obtained through the FM （Factorization Machine） module. Finally， the attention mechanism is used to allocate weights for high-low-order features， and the result of fusion is sent to the output layer to obtain the predicted click-through rate value. Three public datasets Amazon Electronics， Movielens-1m and Taobao User Behavior Set are used to test TDFA. Compared with DeepFM， DIN， DIEN and MIAN， the AUC of TDFA increased by 1.16%，1.51% and 1.10% respectively， and the Logloss decreased by 5.40%，3.51% and 3.73% respectively， which proves the effectiveness of the TDFA model.关键词：TDFA;click-through rate prediction;user history;Transformer;high-low-order features;attention mechanism
- 77
- |
- 27
- |
- 0
发布时间：2023-05-29 - 摘要：The process of EEG acquisition often includes physiological noise and external noise. At present， the external noise can be eliminated by filtering. The usual methods to remove physiological noise are adaptive filter， spatial filter and principal component analysis， but these methods need a certain signal as a prerequisite， and the denoising performance is limited. In recent years， deep learning has been applied to EEG denoising， and the denoising performance has been improved to a certain extent. In the generative adversarial networks， given the advantage that the generated samples gradually approach the real samples in the game between the discriminator and generator. A generated adversarial network model based on the convolutional neural network is designed to eliminate myogenic artifacts and ocular artifacts in EEG signals， and a new loss function is introduced into the Network to make the denoised data more similar to the original data. The designed GAN-1D-CNN model has a correlation coefficient of 0.945 for removing muscle induced noise and 0.894 for removing eye induced noise. The experimental results show that the GAN-1D-CNN model has enhanced its denoising ability for EEG waves and outperforms the existing denoising methods on the benchmark dataset in terms of relevant performance indicators.关键词：electroencephalography;generative adversarial networks;convolutional neural network;EEG denoising;artifact elimination
- 84
- |
- 20
- |
- 0
发布时间：2023-05-29 - 摘要：To reduce the dependence of existing deep named entity recognition（NER） models on high-quality annotated datasets，proposes a named entity recognition strategy based on semi-supervised learning and RoBERTa multi-layer fusion， aiming to boost the performance of Chinese medical named entity recognition. Based on RoBERTa-wwm-ext-BiLSTM-CRF multi-layer representation fusion model， we use pseudo label to expand train set， and build a noise reduction module to mitigate the influence of noise in pseudo label data. We tested our model with the datasets of CCKS 2021 medical named entity recognition task and CBLUE CMeEE， the F1 scores of the proposed method is improved by 1.14% and 1.63% compared with the BERT-BiLSTM-CRF model. The proposed model with semi-supervised learning and RoBERTa multi-layer representation fusion can effectively boost the performance of Chinese medical named entity recognition.关键词：named entity recognition;semi-supervised learning;pre-trained models;deep learning;natural language processing
- 45
- |
- 11
- |
- 0
发布时间：2023-05-29 - 摘要：In order to solve the problems of instability and high risk in predicting natural gas load with a single model， a parallel combination model with meteorological comprehensive factor was proposed. Firstly， Pearson correlation analysis was used to eliminate redundant factors and meteorological comprehensive factors were introduced， then the linear inertia weight of the particle swarm algorithm to improve the algorithm to form the improved particle swarm algorithm and combined with teaching. Finally use the improved algorithm allocated Elman neural network and the weights of NARX neural network to predict the results of parallel combination model. Example analysis shows that the improved PSO has better performance. Compared with single model and average weight model， the prediction accuracy of the parallel combined model with meteorological factors is higher， and the absolute error percentage is improved by 3.03%， and the generalization ability is stronger.关键词：natural gas load forecasting;comprehensive meteorological factors;Elman;NARX;portfolio model
- 20
- |
- 7
- |
- 0
发布时间：2023-05-29 - 摘要：In order to improve the accuracy of face recognition， a classification model of the sparrow search algorithm that integrates the chaotic mapping of Tent is proposed. First of all， improve the sparrow search algorithm， select the test mapping initialization of the sparrow population to improve the quality of the founder's population； add the adaptive adjustment of the inertial weight strategy， enhance the global search ability and convergence speed； Solve the problem of stagnation of algorithms. Then use the improved sparrow search algorithm to adjust the nuclear parameters and punishment parameters of the support vector machine， dynamically adjust the acceptance threshold of the face similarity， and minimize the value of the error classification rate evaluation index. The results show that this method has an accuracy rate of 98.5% in the face recognition classification.关键词：sparrow search algorithm;support vector machine;adaptive adjustment of inertial weight;Cauchy variation;face recognition
- 30
- |
- 13
- |
- 0
发布时间：2023-05-29 - 摘要：Currently， the prevalence of coronary heart disease is increasing year by year， and the trend of youthfulness is obvious. In order to identify the high causative factors of coronary heart disease and improve the diagnostic accuracy， a prediction model was constructed. The medical examination data from Lianyungang Chinese Hospital affiliated with Nanjing University of Traditional Chinese Medicine was selected as the model input， and firstly， the XGBoost coronary heart disease prediction model is constructed， and compared with the support vector machine， logistic regression， random forest and other algorithms； then the Bayesian optimization algorithm was used to optimize the model parameters and obtain the optimal TPE_XGBoost prediction model； finally， the SHAP algorithm was used to analyze and interpret each Finally， the SHAP algorithm was used to analyze and interpret each feature to obtain the high causative factors of coronary heart disease. The results showed that the optimized TPE_XGBoost model performed better in evaluation indexes， with an accuracy of 0.993 7， F1 value of 0.995 5， and AUC of 0.998 3， and the top four risk factors were age， body mass index， low-density cholesterol， and diastolic blood pressure， which provide reference for clinical decision making of coronary heart disease.关键词：coronary heart disease;predictive models;XGBoost algorithm;Bayesian optimization;SHAP interpretation model;feature analysis
- 35
- |
- 5
- |
- 0
发布时间：2023-05-29 - 摘要：The research of sentiment analysis often relies on a large amount of labeled data in specific domain， and domain transfer will lead to the retraining of models. Therefore， cross-domain sentiment analysis tasks emerge as the times require. However， most of the cross-domain sentiment analysis methods mainly consider learning domain-invariant features or perform feature adaptation to deal with the gap between the source domain and the target domain， and do not fully and efficiently utilize the source domain data， so study domain adaptation from the perspective of data transfer and utilization. Unlike previous heuristics that extract transferable samples from the source domain， this paper proposes a general and learnable alternative， gradient data selection （GDS）， to automatically assign weights to samples during training. As a result， non-transferable samples are easily discarded in the sampling process， while transferable samples are sampled more frequently. Related experiments are carried out on two data sets of English and Chinese， and GDS is applied to different models （BERT， RoBERTa， etc.） for generality analysis， the results confirm the effectiveness of this method.关键词：cross-domain sentiment analysis;domain adaptation;data selection;semi-supervised learning;transfer learning;pretrained model
- 20
- |
- 4
- |
- 0
发布时间：2023-05-29 - 摘要：Considering the defects of the low convergence accuracy and the weak exploration ability of standard BAS algorithm， an improved BAS algorithm （IBAS） based on elite opposition-based learning strategy is proposed. Firstly， a beetle swarm is adopted instead of the single individual， and the elite individual of the swarm are used to guide the search path to improve the search efficiency. Secondly， the elite opposition-based learning strategy is introduced to increase the diversity of the population and enhance the exploration ability of the algorithm.To verify the performance of the new algorithm， simulation experiments are conducted on 8 benchmark functions in low dimension and high dimension respectively. The experimental results indicate that the IBAS algorithm is superior to the other three algorithms in terms of optimization accuracy and robustness.关键词：beetle antennae search algorithm;elite individuals;opposition-based learning strategy;global optimization;benchmark functions
- 20
- |
- 7
- |
- 0
发布时间：2023-05-29 - 摘要：With the improvement of living standards， people's requirements for the freshness of fruits and vegetables agricultural products have become higher. Therefore， whether logistics transportation can deliver fruits and vegetables agricultural products within the specified time and ensure the minimum overall cost has become the key. Taking the northwest region as the research object and minimizing the total transportation cost as the optimization objective， taking into account the special quality of fruit and vegetable agricultural products， the environmental requirements of fruit and vegetable agricultural products during transportation， and the time windows of each receiving point. At the same time， adding the required product quantity at the receiving point， setting a penalty time window for each receiving point， and establishing a logistics transportation vehicle path model based on multiple bases under sufficient vehicle conditions， Establish a transportation time cost model based on violation time window penalty during the optimization process. According to the ant colony algorithm， the above constraints are constrained and optimized to obtain an improved ant colony optimization algorithm. Finally， based on the obtained algorithm model， the logistics optimization of the multi base fruit and vegetable agricultural product supply chain in the northwest region was solved. Through single case testing， the path passed by each vehicle and the optimal logistics cost were obtained， verifying the effectiveness of the algorithm. The experimental results show that the improved ant colony optimization algorithm has a good effect on minimum logistics optimization under constraint conditions.关键词：supply chain of agricultural products;logistics optimization;vehicle routing optimization;ant colony
- 13
- |
- 2
- |
- 0
发布时间：2023-05-29 - 摘要：Farmland identification is the premise and foundation to ensure that the automatic operation of agricultural machinery does not cross the boundary. In order to provide environmental information for the automatic operation of agricultural machinery， a farmland identification algorithm based on improved residual network is proposed. The algorithm is improved based on the residual network （ResNet18）， and the channel attention mechanism is added on the basis of the residual structure， which enhances the weight of effective features， reduces the influence of useless features such as noise， and improves the characteristics extraction ability and robustness of the model. Combined with the label smoothing training method and the learning rate cosine annealing decay algorithm， the Silu activation function is used to replace the Relu activation function in the original residual network， which optimizes the convergence effect of the model. The experimental results show that the improved residual network model has a recognition accuracy rate of 98.92% for farmland， grassland， buildings， bare land and desert， which is 3.8% higher than the original ResNet18 model， which proves the superiority of this model.关键词：farmland recognition;feature extraction;residual network;channel attention mechanism;activation function
- 12
- |
- 3
- |
- 0
发布时间：2023-05-29 - 摘要：In response to the problems of cumbersome operation， low intelligence and no suction timing function of traditional aspirators， an intelligent electric aspirator based on OpenHarmony operating system is designed and implemented to meet the suction needs of patients in different scenarios and age groups. The hardware of this instrument is mainly controlled by Huawei Hisi Hi3861， and a peripheral drive circuit is designed to control the operation of the negative pressure pump. A pressure sensor is used to monitor the negative pressure of the gas path in real-time， and an electromagnetic pressure relief valve is used to achieve overpressure protection function. The software is based on the OpenHarmony operating system， running devices through mechanisms such as task scheduling and state machines. It uses a moving average algorithm to preprocess pressure sensor data， and combines threshold algorithm to identify the action of opening the suction tube and time the suction operation. By comparing different threshold algorithms， it is shown that the dynamic dual threshold recognition method has good recognition performance， with a recognition delay of 440ms. The system has high stability and reliability， improves the intelligence level of the electric sputum aspirator， and has broad application prospects.关键词：electric aspirator;suction timing;dynamic double thresholds;air pressure sensors;OpenHarmony
- 20
- |
- 5
- |
- 0
发布时间：2023-05-29 - 摘要：Using inspection images to identify dam defects has some problems， such as strong professionalism， heavy workload and poor timeliness，so a dam defect recognition system based on B/S framework is developed by using the technology of object recognition， transfer learning and other technologies，as well as TensorFlow deep learning framework and Django Web service framework，and 120 crack samples and 85 breakage samples are used for experiments. The experimental results show that a dam defect recognition model with better performance can be trained and the defects in inspection images can be able to recognize with Faster R-CNN and Mask R-CNN pre-training models under the condition of fewer training samples. Under the condition of large number of training samples， reasonable defect classification and accurate defect labeling， the system can train single defect recognition model and multi-defect recognition model with superior performance， which can be used by mobile phones， cameras， drones and other terminals.关键词：dam defect;object recognition;Tensorflow;transfer learning;Django
- 33
- |
- 9
- |
- 0
发布时间：2023-05-29 - 摘要：In order to accurately evaluate the environmental status of fattening pig houses， a remote monitoring system of pig house environment based on the Internet of Things， including a three-layer architecture， was designed. The sensing layer uses PLC as the main controller， and detects the environmental parameters such as temperature， humidity and harmful gas concentration inside the pig house through the shutter box sensor， and displays them on the touch screen. GPRS is used in the transmission layer to realize remote data transmission. The application layer uses Visual Studio to develop the upper computer server software and Android mobile phone client to remotely monitor the piggery environment. At the same time， the fuzzy comprehensive evaluation model of the pigsty is established， the evaluation criteria of the environmental suitability of the pigsty are formulated， the weights of the evaluation factors are determined by the fuzzy analytic hierarchy process， and the ［π］ type function is used as the membership function. Through the fuzzy comprehensive evaluation of the environmental suitability of a pig farm in Yixing City， Jiangsu Province， the results show that the model can provide scientific basis for early warning and control of the pig farm environment.关键词：analytic hierarchy process;Internet of Things;environmental suitability;evaluation system
- 23
- |
- 7
- |
- 0
发布时间：2023-05-29 - 摘要：In order to meet the needs of plant growth and prevent plants from growing in an environment with too high or too low temperature， a miniature plant growth system is designed. The system uses fuzzy control algorithm to control the system temperature， and Jaya algorithm is used to optimize the fuzzy control algorithm to realize the intelligent control of heating electric fan and cooling electric fan， so as to achieve the effect of automatic temperature adjustment. At the same time， this paper designs the supervisory controller to impose the necessary commands on the actuators to ensure the safety of the system under different working conditions. The experimental results show that the improved fuzzy control algorithm has a low overshoot and short adjustment time， as well as good dynamic performance， which basically meets the system control requirements.关键词：Internet of Things;intelligent greenhouse;fuzzy control;Jaya algorithm
- 22
- |
- 5
- |
- 0
发布时间：2023-05-29 - 摘要：Nowadays，many software applications are Web-based applications that run in Internet browsers. These Web applications are very complex， and testing such complex Web applications is difficult. In order to find out program defects， evaluate product quality， and reduce human intervention and repetitive work， an automated testing framework based on Selenium is proposed. Selenium is a powerful set of software tools that work with many browsers， operating systems， programming languages， and testing frameworks， each of which takes a different approach to supporting automated testing of Web applications.The framework facilitates the design and execution of test scripts， thus saving time and cost of manual testing. The framework helps developers analyze their code because of its screenshot feature. The framework can automatically generate test reports for testers. The testing framework makes maintenance and regression testing of new versions of applications very easy.关键词：Selenium;Web;software test;automatic test framework
- 23
- |
- 14
- |
- 0
发布时间：2023-05-29 - 摘要：In modern medical systems， patient data is generally stored in hospital databases and cloud based servers. However， it is very difficult for hospital data maintenance， which may lead to insufficient memory and data security problems. Moreover， patient data is only transmitted and shared in the hospital. When patients are transferred from the hospital， it will lead to treatment lag， greatly added unnecessary trouble. Therefore， based on blockchain and IPFs， this paper proposes a medical big data security model， which uses smart contract to formulate rules to ensure data security， and forms each hospital into an interstellar file system to share and store patient data， which not only solves the problem of insufficient memory， but also solves the problem of data sharing. This scheme is more effective in terms of storage performance and data privacy than existing schemes.关键词：blockchain;smart contract;interstellar file system;data sharing;data security
- 19
- |
- 7
- |
- 0
发布时间：2023-05-29 - 摘要：Aiming at the problem of bidders blindly bidding at high prices in the current sand mining rights auction process， resulting in vicious competition， a blockchain smart contract that integrates credit value and second price sealed auction is designed. Adopting the second price auction mechanism to encourage bidders to bid according to the real estimate， avoiding vicious competition caused by bidders blindly bidding when information is asymmetry. Introduce a credit mechanism to restrain dishonest behaviors such as malicious bidding. Finally， the smart contract was implemented based on the Hyperledger Fabirc1.4 open source consortium chain framework and Golang language， and establish a credible， private data-protecting， fair and third-party online auction environment.Experimental results show that the designed smart contract can effectively restrict bidding， play the role of second-price auction with stable system performance， and meet the daily auction requirements.关键词：second price sealed-auction;credit mechanism;blockchain;smart contract
- 15
- |
- 4
- |
- 0
发布时间：2023-05-29 - 摘要：As a new Internet fundraising and fundraising activity， medical crowdfunding has been widely concerned and applied， the current medical crowdfunding organizations have problems such as misappropriation of funds， unclear fund movements， and leakage of user privacy information. In order to solve the above problems， this paper proposes a medical crowdfunding scheme based on Ethereum smart contracts. Based on the smart contract mechanism and properties， the system structure and business logic are introduced in detail， and the specific system implementation of the platform is given. This paper builds a medical crowdfunding platform based on blockchain technology， which combines blockchain technology with asymmetric encryption to meet the needs of users' privacy not being leaked and the whereabouts of funds being clear. By comparing and analyzing the medical crowdfunding platform and the traditional medical crowdfunding platform， it is proved that the scheme can effectively solve the problems of fund misappropriation， difficulty in returning fundraising and leakage of user privacy.关键词：blockchain;ethereum;smart contract;medical crowdfunding;asymmetric encryption
- 31
- |
- 8
- |
- 0
发布时间：2023-05-29 - 摘要：Current centralized credit banking system has not effectively covered the extensive education and training institutions， nor has it stimulated the initiative of learners to open personal accounts， leading to the fact that credit transfer has not been widely promoted， and its due value in building a lifelong education service system has not been highlighted. To this end， relying on the blockchain's distributed computing framework and data tamper proof and other technical characteristics， we will build a flat chemical sub bank system， create a secure and reliable data system， improve the transparency of the learning achievement certification and credit conversion process， improve users' recognition and satisfaction of credit banks， give full play to the self running advantages of smart contracts， ensure the system's operating efficiency， and promote the construction of credit banks. The distributed system architecture based on the blockchain is an effective means for the credit bank to expand to a wider range of education and training institutions， providing a new idea and referential solution for the construction and promotion of the credit bank.关键词：credit bank;distributed system;Hyperledger Fabric;blockchain;learning outcomes
- 16
- |
- 5
- |
- 0
发布时间：2023-05-29 -
Research on Routing Protocol for Field Observation Instrument Network Based on Unequal Clustering AI导读
摘要：Most of the field observation stations in cold and dry areas are deployed in harsh natural environment， which leads to poor information acquisition ability and seriously hinders the scientific research in these areas. To solve the problem of "hot spots" caused by uneven energy consumption in multi hop routing of field observation instrument network， this paper proposed an energy-balanced unequal clustering routing protocol （EBUCR）. This protocol select cluster head through time， and combine the heterogeneous competition radius to form the unique cluster， the cluster far away from the base station is smaller than the cluster size near the base station. In this way， the energy saved by the cluster head near the base station in the intra-cluster communication can be used to transpond the information of the cluster head far away from the base station. The protocol considers the residual energy of the nodes， the density of the surrounding nodes and the distance from the base station to select the cluster heads， which makes the distribution of cluster heads more reasonable. Meanwhile， the Hybrid Particle Swarm Optimization Algorithm was applied to multi-hop path search， and the optimal path was selected to complete the information transmission among cluster heads， and the communication energy consumption of cluster heads was balanced. The simulation results show that the network lifetime of the EBUCR protocol is 544 rounds， which can effectively extend the network lifetime.关键词：field observation instrument network;unequal clustering;node density;hybrid particle swarm optimization algorithm;network lifetime- 17
- |
- 4
- |
- 0
发布时间：2023-05-29 - 摘要：Aiming at the low efficiency of identity authentication and encrypted data retrieval in the Internet of Things scenario， propose a kind of Internet of Things privacy protection scheme based on trusted identity retrieval. Internet of Things system is mainly composed of remote nodes and central business nodes of the application system， remote node is responsible for performing business， and will produce business data encryption storage， central business node according to the specific requirements of the upper application each time and remote node authentication identity， then request all business related encryption data unified decryption processing， this way the problem of data transmission pressure and low data processing efficiency. This paper uses software defined boundary （SDP） technology to ensure the credibility of access identity， and uses an inverted index based searchable encryption technology to realize the efficient retrieval of cipher text data. First for the remote node encrypted data， the data inverted query index， center business node only need to determine the required data keywords， call retrieval algorithm can obtain the desired encrypted data， to reduce each amount of data transmission network bandwidth pressure， at the same time meet the demand of iot data storage and privacy protection.关键词：Internet of Things;searchable encryption;trusted identity retrieval;software-defined perimeter;information safety
- 29
- |
- 6
- |
- 0
发布时间：2023-05-29 - 摘要：Among the current detection methods for malicious URLs， the detection methods based on the blacklist library have the problem of low timeliness and cannot defend against unknown attacks， while the detection methods based on machine learning rely too much on manually extracted features and require a lot of manpower and time. In response to the above problems， a detection model based on the combination of CNN and XGBoost is proposed， which uses CNN to automatically extract features， classify through XGBoost， and use FWA algorithm to optimize the key parameters of XGBoost. Combined with SVM， logistic regression and other classifications device for comparison. The experimental results show that the model can automatically extract features to achieve active defense， and its accuracy rate reaches 97.2%， which is 3%~5% higher than that of existing detection models.关键词：convolutional neural network;machine learning;malicious URL;FWA;network security
- 10
- |
- 5
- |
- 0
发布时间：2023-05-29 - 摘要：Intelligent reflective surface （IRS） can assist in multi input multi output （MIMO） communication， but its channel estimation is a challenging challenge. Therefore， a segmented estimation scheme utilizing the dual time scale characteristics of the channel is proposed. Due to the fact that the base station （BS） and IRS are usually fixed at a certain position， the channel dimensions between the two are high but slow to change， and can be considered quasi-static. Channel estimation algorithms based on coordinate descent can be used for estimation； However， when the user device （UE） is in a mobile state， the two channel segments of IRS-UE and BS-UE are time-varying but have lower dimensions， which can be estimated using the least squares algorithm， thereby reducing the overall pilot cost. The simulation results show that compared to various existing schemes， this channel estimation scheme reduces pilot overhead by 30% and reduces normalized mean square error by 25dB. It can achieve accurate channel estimation under low pilot overhead conditions.关键词：intelligent reflecting surface;channel estimation;pilot overhead;multiple-input multiple-output;wireless communication
- 13
- |
- 10
- |
- 0
发布时间：2023-05-29 - 摘要：In order to mine the rule of traditional Chinese medicine for treating gall disease and analyze the performance of association rule algorithm on Chinese medicine data， the traditional Apriori algorithm has the characteristics of many candidate item sets， low efficiency and high algorithm complexity. Aiming at this feature， the improved Apriori algorithm based on transaction compression technology and hash technology is introduced. The traditional Apriori algorithm has the characteristics of more duplicate items and longer single transaction in the transaction database of Chinese medicine data， an improved FP-growth algorithm is proposed， which combines these improved association rule algorithms to conduct data mining analysis of traditional Chinese medicine， draw algorithm time efficiency graph， and analyze algorithm efficiency. The results show that the mining results of association rules of each algorithm are consistent， which reflects the basic treatment principle of traditional Chinese medicine in treating gall disease， which is based on the principle of transforming phlegm， eliminating gall and dispersing knots. In terms of algorithm efficiency， the running efficiency of the improved FP-growth algorithm is significantly improved. With the reduction of the threshold of the minimum support or minimum confidence， the running time gap of the algorithm is more obvious.关键词：association rules;Apriori;FP-Growth;frequent itemsets;traditional Chinese medicine;gall disease
- 7
- |
- 2
- |
- 0
发布时间：2023-05-29 - 摘要：The study selected 248 civil transportation airports in China and 228 airlines that have cooperative relationships with them as the dataset， and used a hyper network model construction method to study the hyper network of Chinese airlines from centrality indicators and cumulative probability distribution indicators such as degree， hyper-degree， and hyper-edge hyper-degree. The empirical results show that the hyper network of Chinese airlines exhibits scale-free characteristics， The transportation airports in core cities and a few airlines have higher centrality indicators. The research can provide ideas and development suggestions for the establishment of cooperative network relationship between Chinese airlines and transportation airports.关键词：airlines;hyper-network;scale-free feature;centrality index
- 8
- |
- 3
- |
- 0
发布时间：2023-05-29 - 摘要：With the development of artificial intelligence， more and more enterprises have begun to build intelligent applications. OCR is one of the applications of artificial intelligence， and it is also one of the most popular research directions. In natural scenes， due to the complex shooting environment and uneven quality of shooting equipment， the pictures taken have blurred text， missing text， and complex backgrounds， which will make it difficult to train text recognition models on variable-length datasets. In this case， a new data loading method can accelerate the convergence speed and improve the model recognition accuracy.The method mainly has two parts： ① Design approximate background filling to appropriately increase the proportion of negative samples；②Set the image scaling threshold to ensure the shape of the real data and appropriately scale the data to achieve text data in natural scenes style. Compared with the classical and commonly used data loading methods， the data loading method proposed in this paper has faster fitting speed and recognition accuracy.关键词：text recognition;text blur;data loading;image scaling threshold;approximate background fill
- 11
- |
- 6
- |
- 0
发布时间：2023-05-29 -
Tobacco Image Grading Method Based on Densely Connected Convolutional Networks and Hybrid Attention AI导读
摘要：In view of the decline of classification accuracy caused by the influence of environment and personal status in the actual purchase process of tobacco leaves，a deep learning model combining dense connected convolutional neural network and hybrid attention mechanism is proposed.By improving the residual attention network，the model use densely connected convolution module to replace the main branch residual module in the original network attention module and multiple residual modules in the network output stage， enhance the learning performance of branch features， reduce the amount of parameters and alleviate the problem of gradient disappearance；At the same time， the two attention mechanism modules in the residual attention network are used， and by adding spatial attention mechanisms （SAM） before the two attention mechanism modules， the information of tobacco leaf feature map is weighted and extracted in the two dimensions of space and channel， so as to obtain more comprehensive feature information.Our network not only reduces the network depth， but also improves the speed and recognition accuracy. The experimental analysis of 5 000 tobacco leaf images in 10 grades shows that the classification accuracy of the network is improved by 8.19% compared with the original network， 7.72% compared with VGGnet19， the amount of network parameters is reduced by 45% compared with ResNet34， and the training speed is improved by 38.11%. It can be seen that this method has better recognition effect and faster recognition speed for different grades of tobacco leaves， and can be well applied to tobacco leaf classification in tobacco production.关键词：tobacco grading;image classification;hybrid attention;densely convolutional network;deep learning- 19
- |
- 8
- |
- 0
发布时间：2023-05-29 - 摘要：Fast and efficient detection of missing and defective fasteners on high-speed railways is of great significance to ensure the maintenance and operational safety of railway track lines. A high-speed railway fastener defect detection and extraction algorithm based on improved Faster R-CNN is proposed using object detection technology， which can significantly improve detection speed and accuracy compared to manual inspection and static detection. Firstly， using ResNet-101 network instead of VGG-16 network， combined with FPN to extract image features and perform feature map fusion. Then， ROI Align was used instead of ROI Pooling to avoid double quantization and eliminate the bounding box bias of the Faster R-CNN model itself. Finally， the Soft-NMS algorithm is used to replace the traditional NMS algorithm to suppress non maximum values and improve detection efficiency. The experiment shows that this method improves the accuracy of extracting and detecting railway fastener defects by 10.58% compared to the traditional Faster R-CNN method， and has certain practical value.关键词：target detection;Faster R-CNN;fastener defect;ResNet-101;ROI Align;Soft-NMS
- 14
- |
- 4
- |
- 0
发布时间：2023-05-29 - 摘要：Aiming at the problem of domain offset between different camera views， such as visual differences caused by changes in domain information such as occlusion， illumination， background， etc.， a cross-view pedestrian recognition algorithm based on joint dictionary pair learning is proposed. Specifically， because images from the same camera perspective have domain similarity， the dictionary used to represent domain information is refined through low-rank decomposition， that is， one dictionary represents domain information from the same camera perspective， and another dictionary represents pedestrian appearance feature information. The main idea of this method is to make all pedestrian images from the same camera perspective have the same sparse representation in the dictionary， so as to separate the domain information and pedestrian feature information shared by pedestrian images from each perspective. Experiments show that the Rank-1 value of the proposed method is better than that of the suboptimal algorithm in PRID2011， CUHK01 and i_ The LIDS data set is improved by 1.2%， 1.8% and 4.49% respectively， and the recognition performance and robustness are better， so as to provide reference and reference for cross-view pedestrian recognition.关键词：person re-identification;joint dictionary pair learning;domain information separation;low-rank decomposition
- 17
- |
- 6
- |
- 0
发布时间：2023-05-29 -
Construction and Practice of Postgraduate Artificial Intelligence Course Based on Open Source Mode AI导读
摘要：The rise of open source platforms has provided impetus for the development and continuous innovation of artificial intelligence. By analyzing the problems， challenges， and requirements of the times in the construction of current graduate artificial intelligence courses， and proposes principles and implementation paths for embracing open source platforms in the construction of artificial intelligence courses. By focusing on analyzing and designing an open-source capability evaluation system， we aim to drive innovation in course construction through open-source evaluation， enhance students' comprehensive abilities in research in the field of artificial intelligence， and cultivate research-oriented talents familiar with open-source.关键词：open source mode;talent training;artificial intelligence;curriculum construction- 21
- |
- 12
- |
- 0
发布时间：2023-05-29 - 摘要：The UNESCO Guide for Policy Makers on Artificial Intelligence and Education introduces the necessary knowledge for policy makers on artificial intelligence， analyzed the development opportunities of artificial intelligence for education， assessed the risks and challenges of artificial intelligence and education， and proposed recommendations for AI education regulation，interpreting it helps promote and regulate the deep integration of artificial intelligence and education. This guide prompts that we should explore the paradigm of artificial intelligence education governance， improve the top-level design of education governance， and ensure the fairness and ethical standards of artificial intelligence education； Strengthen humanistic care and deliberation on human intelligence， promote artificial intelligence technology to empower education， and build a new education ecosystem； Harness the literacy level of artificial intelligence to ensure the achievement of sustainable development goal 4 in education，that is ensure inclusive and equitable quality education， and enable all people to enjoy lifelong learning opportunities.关键词：UNESCO;artificial intelligence education;deep integration
- 13
- |
- 11
- |
- 0
发布时间：2023-05-29 - 摘要：In response to society's demand for software talents， propose reform and optimisation of the curriculum system of software engineering majors based on the teaching model of integrated undergraduate and postgraduate courses， basing on the analysis of the existing training models of universities. This model integrates the teaching contents of undergraduate and graduate students into one whole according to the knowledge framework of software practitioners， and carries out hierarchical training according to different talent standards. The implementation of this model will have a positive effect on the cultivation of complex software talents who meet the needs of society.关键词：software talents;integration of undergraduate and postgraduate;curriculum reform;teaching mode
- 14
- |
- 7
- |
- 0
发布时间：2023-05-29 - 摘要：Aiming at the problems of single teaching methods， insufficient depth of teaching content， and single teaching evaluation and evaluation in traditional intelligent robot courses， an online and offline multi-hybrid teaching mode is proposed. This model includes promoting the combination of online and offline training in teaching methods， building a hierarchical mixing of teaching content and process， and establishing a multi-dimensional mixed assessment in teaching evaluation.Clarify the objectives and content of curriculum construction， and introduce the progress and achievements of curriculum construction from three aspects： the construction of mixed teaching resources， the reform of teaching content and process， and the implementation of multi-dimensional mixed assessment. After more than two years of reform and practice， the online and offline multi-hybrid teaching mode of intelligent robotics courses has achieved positive results， and promoted the cultivation of outstanding engineers in new engineering.关键词：new engineering;intelligent robot;blended teaching;course reform
- 14
- |
- 10
- |
- 0
发布时间：2023-05-29 - 摘要：In order to effectively and organically integrate ideological and political education and professional education， and at the same time cultivate students' patriotism， to achieve the goal of promoting teaching reform， improving teaching effectiveness， and cultivating qualified talents， take the data mining course as an example， through analyzing the problems in teaching， this paper proposes a plan to reform the ideological and political supply side from three dimensions of teaching philosophy， classroom cases， and teaching evaluation. The results of teaching reform practice indicate that students' professional knowledge and ability levels have improved， and their comprehensive qualities have improved. This reform plan has better achieved the goal of talent cultivation and provided reference for the integration of other courses with ideological and political education.关键词：ideological and political education;data mining;supply-side reform
- 17
- |
- 7
- |
- 0
发布时间：2023-05-29 -
Teaching Exploration of Software Process Management Course for Application-oriented Universities AI导读
摘要：With the progress of the information society and the development of the software industry， software enterprises have increased the demand for the process management ability of professional developers and testers. In order to explore how the undergraduate teaching of software engineering in application-oriented universities can meet the needs of software enterprises for talents and cultivate application-oriented talents for enterprise practice. Based on the organizational logic of individual developers， small development teams and multi-team collaboration， the theory and practical teaching content of software process management for enterprise application design is designed to build a software process management course oriented by talent demand. First， divide the software process management model into traditional and modern enterprise applications， set the core， expand the teaching content， and realize the close combination of traditional process management methods and enterprise practice. Then， when implementing the expanded teaching content， the flexible teaching plan for personalized education is adopted to solve the problems of tense class hours and diversified career planning objectives. The actual teaching shows that the proposed teaching design scheme can effectively promote the connection between software engineering undergraduate training and industry， shorten the gap between college education and software enterprise talent demand， improve the undergraduate software process management ability， and provide a new method for the software process management course teaching in application-oriented universities.关键词：software process management;personalized education;process management capability;computer education- 16
- |
- 12
- |
- 0
发布时间：2023-05-29 - 摘要：Knowledge graph is an old and new topic，a new form of knowledge engineering in the new era，and one of the important directions of artificial intelligence.The application of knowledge graph in the medical field has played a great role in pushing the medical data and the treatment process between doctors and patients to a certain extent.The feature of this paper is the construction of medical knowledge graph，from seven aspects：medical knowledge data source，medical knowledge extraction，medical knowledge fusion，medical knowledge representation，medical knowledge storage，medical knowledge reasoning，and medical quality assessment，and then summarizes the application of medical knowledge graph in clinical decision support diagnosis，medical knowledge popularization，medical question answering system and intelligent semantic search，and finally discusses the current challenges faced by medical knowledge graph，in order to provide reference for the development and construction of medical knowledge graph.关键词：medical knowledge graph;construction technology;artificial intelligence;big data
- 14
- |
- 10
- |
- 0
发布时间：2023-05-29 - 摘要：Many practical engineering optimization problems usually involve optimizing multiple conflicting objectives at the same time， and the calculation of the objective function mainly relies on time-consuming simulation experiments， such problems can generally be called expensive multiobjective optimization problems. Surrogate-assisted evolutionary algorithms use machine learning methods to build surrogate models and assist algorithms for evaluation， which makes surrogate-assisted evolutionary algorithms a popular method to solve such problems. According to the scale of the problem， the relevant algorithms are divided into two categories， the characteristics of each type of problem are described， the related algorithms are classified and sorted out， and the advantages and disadvantages of each algorithm are explained， so that people can intuitively understand the research progress of continuous expensive multiobjective optimization problems and better carry out follow-up research work.关键词：multiobjective optimization;expensive multiobjective optimization;surrogate-assisted evolutionary algorithm;surrogate model;machine learning
- 11
- |
- 6
- |
- 0
发布时间：2023-05-29
- 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