Search

With the development of highways, there has been an increase in toll evasion by unscrupulous individuals who employ various means to avoid paying fees. This has made it essential for highway management to focus on effective methods to prevent and manage toll evasion, as well as regulate toll-related operations. However, traditional auditing techniques are heavily reliant on manual toll data and lack sufficient auxiliary data such as video and pictures, making big data analysis and in-depth data analysis difficult. Consequently, these methods can only identify a portion of evasion behavior, with a significant number of potential evasion cases remaining unidentified. To address this issue, we propose utilizing vehicle re-identification to enable cross-camera identification and continuous tracking of vehicles, allowing for accurate calculation of their operating mileage. However, vehicle re-identification using images captured by highway cameras is a challenging task, primarily due to the low resolution of the images and the lack of relevant datasets. To tackle these problems, we construct a large-scale, low-resolution dataset based on images captured by cameras on highways. We also propose a Part-guided Feature-combination Network (PFNet) to analyze the dataset. PFNet adopts a three-stage approach, wherein each stage of the network requires independent training. In the first stage, the Pyramid Scene Parsing Network (PSPNet) is improved to segment the entire vehicle into parts. In the second stage, based on a comparative learning approach, the Siamese Feature Extraction Module (SFEM) module was created to extract features of pairs of vehicle parts. To train the SFEM, pairs of positive samples, which consist of the same vehicle component pairs captured by different cameras, and pairs of negative samples, which consist of different vehicle component pairs are needed. To ensure that the positive sample pairs are closer to each other in the embedding space, and the negative samples are as far away as possible from each other, contrast loss is utilized. Moreover, The SFEM incorporates VGG16 and a Multi-period Convolutional Block Attention Module (MCBAM), which enhances the learning ability of the conventional Convolutional Block Attention Module (CBAM). In the third stage, information propagation among the vehicle parts is achieved using a Multi-layer Perceptron (MLP), which calculates the similarity after combining the features of all parts. Our self-built test set demonstrates that PFNet is effective for continuous monitoring of vehicles on highways under multiple cameras, resulting in improved accuracy of vehicle re-identification, with Top1 accuracy of 93.61% and Top5 accuracy of 99.44%. [ABSTRACT FROM AUTHOR]

In the era of information and communication technology (ICT), the advancement of science and technology has led to a trend of diversification in map representation. However, the lack of professional knowledge means that there is still a challenge in determining the appropriate type of thematic map for land use expression. To address this issue, this paper proposes a knowledge recommendation method for land use thematic maps based on the theory of visualization dimensions. Firstly, we establish a knowledge ontology of land use thematic maps centered on spatial data, data characteristics, visualization dimensions, thematic map forms, and application scenarios. A land use thematic map knowledge graph is constructed through knowledge extraction and storage operations. Secondly, knowledge embedding is performed on the knowledge graph to enable the knowledge-based expression of map visualization elements. Finally, based on the knowledge elements of land use thematic expression, a similarity calculation model is established to calculate the similarity between input data and the spatial data characteristics, visualization dimensions, and application scenarios within the knowledge graph, deriving a comprehensive similarity result to achieve precise recommendation for land use thematic map forms. The results show that the method can provide a more accurate visualization reference for the selection of land use themes, meeting the diversified needs of land use thematic expression to a certain extent. [ABSTRACT FROM AUTHOR]

Geometric constraint solvers are crucial for computer-aided design (CAD), and their algorithms are the focus of research. Current geometric constraint solvers based on traditional numerical methods lack support for multi-solution problems, so we propose a hybrid algorithm that combines the genetic algorithm, which is good at global convergence, and Powell’s method, which is good at local refinement, to address the limitations of traditional numerical methods in geometric constraint solving algorithms (sensitivity to initial values, susceptibility to falling into local optimums, and being only able to obtain a single solution) and the challenges of intelligent optimization algorithms (complex parameter tuning, slow convergence and low accuracy). Our method has a large accuracy improvement over the comparison method in basically all test cases, and its effciency can also meet the needs of real geometric constraint solving scenarios. This research provides new insights into the design of geometric constraint solving algorithms, offers a fresh perspective on improving the performance and generality of solvers, and contributes to technological advances in the CAD field. [ABSTRACT FROM AUTHOR]

With the advent of the era of big data, an increasing amount of duplicate data are expressed in different forms. In order to reduce redundant data storage and improve data quality, data deduplication technology has never become more significant than nowadays. It is usually necessary to connect multiple data tables and identify different records pointing to the same entity, especially in the case of multi-source data deduplication. Active learning trains the model by selecting the data items with the maximum information divergence and reduces the data to be annotated, which has unique advantages in dealing with big data annotations. However, most of the current active learning methods only employ classical entity matching and are rarely applied to data deduplication tasks. To fill this research gap, we propose a novel graph deep active learning framework for data deduplication, which is based on similarity algorithms combined with the bidirectional encoder representations from transformers (BERT) model to extract the deep similarity features of multi-source data records, and first introduce the graph active learning strategy to build a clean graph to filter the data that needs to be labeled, which is used to delete the duplicate data that retain the most information. Experimental results on real-world datasets demonstrate that the proposed method outperforms state-of-the-art active learning models on data deduplication tasks.

[Display omitted] • A directed weighted graph is established based on disaster-prone environment. • A balanced non-landslide sampling is designed based on similar environments. • A graph neural network model is proposed for landslide susceptibility assessment. Landslide susceptibility assessment (LSA) is vital for landslide mitigation and management. Existing LSA methods only consider local environmental characteristics associated with landslides, while neglecting the characteristics of remote yet interconnected geographic environments. In addition, the datasets for landslide samples do not consider the equilibrium among various geographic environmental characteristics. Ultimately, the landslide susceptibility mapping (LSM) generated is unreliable. We propose an LSA method based on the similarity of geographic environments and the equilibrium of landslide samples. We selected the Bailong River Basin (BRB) as the study area. Our proposed LSA method has four steps: (1) We constructed slope units using the curvature watershed method (CWM) based on the digital elevation model (DEM) to generate geographic nodes and selected 12 landslide influencing factors (LIFs) through multicollinearity analysis to extract geographic node features. (2) We constructed an association graph of geographic environment similarity constraints based on geographic node features using the cosine similarity method to correlate landslide and geographical environments. (3) We constructed sample datasets for an equal number of landslide and non-landslides based on various geographic environments. (4) We constructed a graph neural network (GNN) model based on the association graph to predict the susceptibility of unknown geographic nodes, obtaining a reliable LSM. Compared with random forest (RF), support vector machine (SVM), and multi-layer perceptron (MLP) models, our proposed model framework outperforms in the LSA. The area under the receiver operating characteristic curve (AUC) values improved by 6.2%, 5.7%, and 1.8%, respectively. Specifically, our model can predict landslides across space in areas with few samples. Our proposed model offers an effective approach and essential technical support for landslide disaster investigation and prevention. [ABSTRACT FROM AUTHOR]

Since the identification of microRNAs (miRNAs), empirical research has demonstrated their crucial involvement in the functioning of organisms. Investigating miRNAs significantly bolsters efforts related to averting, diagnosing, and treating intricate human maladies. Yet, exploring every conceivable miRNA–disease association consumes significant resources and time within conventional wet experiments. On the computational front, forecasting potential miRNA–disease connections serves as a valuable source of preliminary insights for medical investigators. As a result, we have developed a novel matrix factorization model known as Hessian-regularized L 2 , 1 nonnegative matrix factorization in combination with deep learning for predicting associations between miRNAs and diseases, denoted as H R L 2 , 1 -NMF-DF. In particular, we introduce a novel iterative fusion approach to integrate all similarities. This method effectively diminishes the sparsity of the initial miRNA–disease associations matrix. Additionally, we devise a mixed model framework that utilizes deep learning, matrix decomposition, and singular value decomposition to capture and depict the intricate nonlinear features of miRNA and disease. The prediction performance of the six matrix factorization methods is improved by comparison and analysis, similarity matrix fusion, data preprocessing, and parameter adjustment. The AUC and AUPR obtained by the new matrix factorization model under fivefold cross validation are comparative or better with other matrix factorization models. Finally, we select three diseases including lung tumor, bladder tumor and breast tumor for case analysis, and further extend the matrix factorization model based on deep learning. The results show that the hybrid algorithm combining matrix factorization with deep learning proposed in this paper can predict miRNAs related to different diseases with high accuracy. [ABSTRACT FROM AUTHOR]

As smart homes become more prevalent, users expect to control smart devices through natural language commands and desire personalized smart home services. However, existing challenges include the interoperability of smart devices and a comprehensive understanding of the user environment. To address these issues, this paper proposed a framework supporting personalized smart home service recommendations for device-end users. Firstly, it constructed a runtime knowledge graph to reflect contextual information in specific smart homes and generated scenario-based sentences. Secondly, it trained a general recommendation model using pre-collected natural language instructions and corresponding scenario-based sentence representations from users in common scenarios. Finally, users interacted with smart home devices and services through natural language on the device end while fine-tuning the weights of the general model through feedback to obtain a personal model. Experimental results on three datasets-basic instruction set, paraphrase set, and scenario instruction set show that the personal model achieves an accuracy improvement of 6.5% to 30% compared to word embedding methods and 2.4% to 25% compared to the Sentence-BERT model, which validates that the device-end deep learning-based smart home service framework has a high service recommendation accuracy and effectively manages smart home devices and services. [ABSTRACT FROM AUTHOR]

• The ADNMT uses two asymmetric networks to generate pseudo-labels for each other by clustering and iterative training. • Based on alternate training and mutual teaching to optimize the pseudo-labels. • The SCIC compensates the similarity distance of images captured by inter-camera. • The SSIC further suppresses the similarity distance by calculate the similarity distribution. Unsupervised person re-identification (Re-ID) has always been challenging in computer vision. It has received much attention from researchers because it does not require any labeled information and can be freely deployed to new scenarios. Most unsupervised person Re-ID research studies produce and optimize pseudo-labels by iterative clustering algorithms on a single network. However, these methods are easily affected by noisy labels and feature variations caused by camera shifts, which will limit the optimization of pseudo-labels. In this paper, we propose an Asymmetric Double Networks Mutual Teaching (ADNMT) architecture that uses two asymmetric networks to generate pseudo-labels for each other by clustering, and the pseudo-labels are updated and optimized by alternate training. Specifically, ADNMT contains two asymmetric networks. One network is a multiple granularity network, which extracts pedestrian features of multiple granularity that correspond to numerous classifiers, and the other network is a conventional backbone network, which extracts pedestrian features that correspond to a classifier. Furthermore, because the camera style changes seriously affect the generalization ability of the proposed model, this paper designs Similarity Compensation of Inter-Camera (SCIC) and Similarity Suppression of Intra-Camera (SSIC) according to the camera ID of the pedestrian images to optimize the similarity measure. Extensive experiments on multiple Re-ID benchmark datasets show that our proposed method achieves superior performance compared with the state‐of‐the‐art unsupervised person re-identification methods. [ABSTRACT FROM AUTHOR]

Copyright of Electric Power is the property of Electric Power Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Aiming at the problem of the remaining useful life prediction accuracy being too low due to the complex operating conditions of the aviation turbofan engine data set and the original noise of the sensor, a residual useful life prediction method based on spatial–temporal similarity calculation is proposed. The first stage is adaptive sequence matching, which uses the constructed spatial–temporal trajectory sequence to match the sequence to find the optimal matching sample and calculate the similarity between the two spatial–temporal trajectory sequences. In the second stage, the weights of each part are assigned by the two weight allocation algorithms of the weight training module, and then the final similarity is calculated by the similarity calculation formula of the life prediction module, and the final predicted remaining useful life is determined according to the size of the similarity and the corresponding remaining life. Compared with a single model, the proposed method emphasizes the consistency of the test set and the training set, increases the similarity between samples by sequence matching with other spatial–temporal trajectories, and further calculates the final similarity and predicts the remaining use through the weight allocation module and the life prediction module. The experimental results show that compared with other methods, the root mean square error (RMSE) index and the remaining useful life health score (Score) index are reduced by 12.6% and 14.8%, respectively, on the FD004 dataset, and the RMSE index is similar to that in other datasets; the Score index is reduced by about 10%, which improves the prediction accuracy of the remaining useful life and can provide favorable support for the operation and maintenance decision of turbofan engines. [ABSTRACT FROM AUTHOR]

In the era of information and communication technology (ICT), the advancement of science and technology has led to a trend of diversification in map representation. However, the lack of professional knowledge means that there is still a challenge in determining the appropriate type of thematic map for land use expression. To address this issue, this paper proposes a knowledge recommendation method for land use thematic maps based on the theory of visualization dimensions. Firstly, we establish a knowledge ontology of land use thematic maps centered on spatial data, data characteristics, visualization dimensions, thematic map forms, and application scenarios. A land use thematic map knowledge graph is constructed through knowledge extraction and storage operations. Secondly, knowledge embedding is performed on the knowledge graph to enable the knowledge-based expression of map visualization elements. Finally, based on the knowledge elements of land use thematic expression, a similarity calculation model is established to calculate the similarity between input data and the spatial data characteristics, visualization dimensions, and application scenarios within the knowledge graph, deriving a comprehensive similarity result to achieve precise recommendation for land use thematic map forms. The results show that the method can provide a more accurate visualization reference for the selection of land use themes, meeting the diversified needs of land use thematic expression to a certain extent.

Tourism marketing strategy is the top priority for all tourist attractions and tourism-related enterprises, and how innovative tourism marketing strategy in the digital era has become the key to improving the competitiveness of attractions and enterprises. This paper uses data mining technology as a basis to study feasible solutions for innovative tourism marketing strategies for scenic spots and enterprises under data-driven. A formalized label representation is used to construct an extended user portrait conceptual model by clustering the tourists’ contextual information using K-means. Tourism marketing targets users based on similarity calculation results, and association rules are used to mine the alignment association rules between tourists’ contextual features and attractions. Combined with the big data framework system Hadoop technology, we have built a tourism marketing platform that is used to analyze the tourism preferences of different tourists and formulate effective tourism marketing strategies. This paper, based on the data-driven tourism marketing strategy, can significantly improve scenic spot traffic. After marketing the Biluo Tower scenic spot in October, the average value of the traffic is up to 20,384 people, compared with the pre-marketing increase of 56.71%. At the same time, the marketing strategy significantly improves the satisfaction of tourists, and after the implementation of the marketing strategy, the satisfaction of tourists in the scenic area generally improves to more than 0.975.

With the rapid development of information technology, how to enable teachers and students to quickly find and filter information of interest in massive collections has become a hot issue for many scholars. This paper enhances the traditional collaborative filtering algorithm by utilizing the knowledge-sharing model. Specifically, we calculate similarity using keyword information from items and dynamic information from users based on similarity calculations related to item features and user attributes. The relevant information about users and items is fully utilized to successfully alleviate the problem of new items and new users, and the entire process of collaborative filtering recommendations for libraries is designed. The improved collaborative filtering-based algorithm can achieve a recommendation accuracy of more than 60% and recommend more accurate books to users. The average recommendation rate of the book recommendation algorithm is 0.056 higher than the average recommendation rate of other recommender systems, indicating a higher recommendation rate that can better match the needs of users and alleviate the cold-start problem to a certain extent.

Man and nature are interconnected and interdependent. The naturalization of man reflects the scientific nature of the landscape, the development of ecotourism is to establish the development of man and nature in harmony and unity of the development of ideas. This paper proposes measures to enhance the tourism experience based on the above concept. After a series of operations, such as similarity calculation and data cleaning, the user category, consumption level, and geographic location preference are calculated to establish a user model and portrait based on behavioral analysis. Based on LASSO penalty regression, the link between covariates is incorporated into the penalty term to complete the optimization and improvement of the LASSO regression model. Associate the user preference degree model with the tourism demand prediction model based on user preference. The model is applied in practice, and the average error rate for the number of tourists predicted by the LASSO regression coefficient is about 9.9%, which is smaller than the average error rate of the comparison model. Based on the results of the demand prediction, strategies related to the development of sports tourism projects are proposed, and the impact of sports tourism projects on the environment is evaluated. Based on the IPA analysis method, the user’s tourism experience of the subject area is assessed, and the user is more concerned about the internal transportation of the tourist place, with I and P values of 12.455 and 1.457, respectively, and the performance effect of the subject area in this point is poor, which needs to be solved in time.

Combining the user model and similarity calculation, a personalized recommendation method based on the user preference model is proposed, and the user model is adjusted to update the overall framework of personalized recommendation of user reading preference in response to the user’s operational behaviors such as collecting, commenting, reading, and annotating. We analyze and organize the knowledge service interface structure of the social reading platform and the knowledge service mechanism model based on users’ cognitive, emotional, and social needs and clarify the whole process of satisfying users’ reading needs. According to the age, occupation, and regional characteristics of the users of the social reading platform, as well as the reading interest labels of the users, we explore the influence of different user behaviors on the personalized recommendation results of the social reading platform and the recommendation website, and the more the optimal value of the parameter tends to be 1, the greater the influence of the social reading platform on the recommendation results. The satisfaction index of the social reading platform is 68.50807. The current social reading is not good enough to meet the needs of users. It should be optimized for reading behavior, reading skills, reading knowledge, and reading environment for knowledge service recommendation, respectively.

This paper applies the cluster analysis method to classify different Civics contexts in university English courses, and improves the accuracy of Civics context text clustering by extracting the word stems of English Civics context texts. The text similarity calculation is used to remove feature words from the preprocessed text, to converge into a collection of feature words. The texts containing English Civic and Political Context are clustered and analyzed, and the clustering effect of Civic and Political Context texts is judged according to the clustering evaluation. The status quo of Civic and Political integration into English curriculum is explored from different dimensions, and corresponding optimization strategies are proposed with the converged data. The results show that the Civic-Political contexts of English courses are mainly divided into three categories, and based on the clustering of the Civic-Political contexts of university English courses, the contexts of “living, learning and doing” are relatively independent. The contexts “history, culture and society” and “science and technology” are closer. In integrating Civics and Politics into English teaching, the mean value of the dimension of teaching method is 4.212, which is in a better integration situation.

In this paper, we carry out an innovative reform of architectural design professional courses in colleges and universities, focusing on improving the effect of practical teaching by combining knowledge maps and new technological means. Aiming at the problems of long practice cycle, high cost and difficulty verifying the effect in traditional teaching, it promotes the transformation of teaching mode to immersion, interaction and autonomy. The article discusses the application of knowledge maps in optimizing the course content, improving the effect of teaching organization, such as the association model of knowledge elements and learning resources, and calculating the ontological conceptual similarity of knowledge nodes. Meanwhile, the introduction of augmented reality technology and the Analysis of the correlation between learning styles and academic performance highlight the critical role of technology in enhancing the efficiency and quality of learning. The difference in average grades before and after the teaching reform in the experimental class was significant, with the most enormous difference of 10.76 points in the second round of test results. The most significant improvement was observed in the knowledge part of the course, with an increase of 0.72, and in the practical part, with an increase of 0.39. The students’ architectural design and practice skills were significantly improved in all dimensions. The teaching reform program proposed in this paper effectively enhances the architectural design program’s teaching quality and students’ practical ability.

In this paper, based on the practical framework of teaching folk dance in colleges and universities under the background of big data, the spatial and temporal characteristics of folk dance are sorted out, and the DTW algorithm is proposed to calculate the similarity of folk dance movements, and a model for estimating the gesture of normative movements of folk dance is constructed by combining multiscale feature fusion. Six kinds of folk dance movements were collected to establish the data samples of folk dance movements in order to verify the effectiveness of the calculation method of dance movement similarity and to obtain the results of the teaching of folk dance normative movements. Analyze the content of dance movements and the dance normative movement gesture estimation method to explore the expressiveness of folk dance movements. To research the effectiveness of emotional dance movement teaching in teaching folk dance in two dimensions, three segments of folk dance were chosen. The data show that the second segment of the dance, M=9.45, SD=2.15, was improved to M=11.55, SD=3.16. Affective dance movement teaching has a significant impact on the teaching of folk dance in colleges and universities, with improvements in both dance movement and dance expression.

At present, although colleges and universities are actively exploring the construction of innovative and entrepreneurial teaching classrooms, they do not have the expected effect in practice. In this paper, based on the compressed perception technology, the complex data in the multivariate data chain network is sparsely represented, and multiple linear subsets in the data matrix are calculated by similarity. Statistical inference is used to generate the recommendation module after describing the encoding of innovation and entrepreneurship information packages in the data. Acquire the characteristics of students’ interest in innovation and entrepreneurial learning, create an interest graph module, and integrate the multi-perspective attention network to overcome the issue of recommendation bias. The analysis of teachers’ competence and students’ learning effectiveness involves the use of empirical testing methods. The results showed that among the 20 teachers, the teacher numbered 11 had a good performance with a competency of 0.8632 on entrepreneurship resources. The students’ 4 dimensions of innovation and entrepreneurship competence improved by more than 2 points, and the standard deviation was within the acceptable range. In the effect of the teaching model application, the mean value of students’ scores after improvement is more than 28. Teachers’ competence and students’ innovation and entrepreneurship ability are improved according to the model constructed in this paper.

This paper presents a semantic analysis of corpus-assisted collocation teaching based on the random matrix theory, taking into account the characteristics of the computer corpus. The construction of a word-document matrix is based on the potential semantic analysis of the corpus and the intrinsic relationships between words and documents. The aggregation and combination relations of collocation, i.e., the lexicon of word meanings, are used to calculate affirmative sentence similarity, incorporating dependent syntactic and semantic analysis. The error analysis method was applied to the teaching of collocations to analyze the common types of errors and the variability of writing errors in English writing and to discuss the correlation between the performance of English composition and the number of collocations used. The correlation coefficient between composition scores and the number of collocations in compositions was 0.924, and the composition scores of college students had a positive linear relationship with the number of collocations.

An improved Hybrid Collaborative Filtering algorithm (H-CF) is proposed, addressing the issues of data sparsity, low recommendation accuracy, and poor scalability present in traditional collaborative filtering algorithms. The core of H-CF is a linear weighted hybrid algorithm based on the Latent Factor Model (LFM) and the Improved Item Clustering and Similarity Calculation Collaborative Filtering Algorithm (ITCSCF). To begin with, the items are clustered based on their attribute dimension, which accelerates the computation of the nearest neighbor set. Subsequently, H-CF enhances the formula for scoring similarity by penalizing popular items and optimizing unpopular items. This improvement enhances the rationality of scoring similarity and reduces the impact of data sparseness. Furthermore, a weighting function is employed to combine the various improved algorithms. The balance factor of the weighting function is dynamically adjusted to attain the optimal recommendation list. To address the real-time and scalability concerns, the algorithm leverages the Spark big data distributed cluster computing framework. Experiments were conducted using the public dataset MovieLens, where the improved algorithm's performance was compared against the algorithm before enhancement and the algorithm running on a single machine. The experimental results demonstrate that the improved algorithm outperforms in terms of data sparsity, recommendation personalization, accuracy, recall, and efficiency. [ABSTRACT FROM AUTHOR]

Recommendation performance in collaborative filtering is significantly influenced by the selected similarity measure. The Jaccard similarity measure which considers the number of co-rated items between users has been widely used in CF owing to its good performance and simplicity. However, it does not consider rating information during similarity estimation. Therefore, several studies have been conducted to address the limitations of Jaccard similarity and Rating_Jaccard, which explains that the number of items identically co-rated by two users is one of them. However, Rating_Jaccard still suffers from certain shortcomings: it cannot measure the similarity of many user pairs and provides low similarity values as the number of co-rated items increases. Therefore, this study proposes new similarity measures to address these issues. The recommendation performances of the proposed similarity measures were evaluated on six datasets widely used in recommendation systems: CiaoDVD, FilmTrust, MovieLens100K, MovieLens1M, Amazon, and Netflix. The experimental results show that, on average, the proposed measures outperformed existing alternatives, including Rating_Jaccard. [ABSTRACT FROM AUTHOR]

To solve the problem that existing algorithms were difficult to effectively identify sensitive sequences in genomic data for long-read with high error rate, a recognition algorithm using hybrid CPU and GPU parallel computing, called CGPU-F3SR, was proposed.Firstly, the long-read in genomic data were partitioned into multiple short-read, and the Bloom filtering mechanism was used to avoid repeated calculation of the short-read.Secondly, the k-mer coding strategy was used to extract in parallel the error information of all short-read, the recognition accuracy was promoted by improving the sequence similarity calculation model.Finally, CPU and GPU were used to coordinate and parallel to accelerate the calculation of short-read similarity to improve recognition efficiency.As a result, both two types of sensitive sequences including short tandem repeats and disease related sequences could be identified efficiently and accurately from genome data for long-read with high error rate.The experimental results of recognizing sensitive sequences from genomic data for long-read with length 100～400 kbp each show that, compared with existing parallel algorithm, the average recognition accuracy and precision rate of proposed CPU/GPU parallel algorithm CGPU-F3SR are increased by 7.77% and 43.07% respectively, its average false positive rate is reduced by 7.41%, and its average recognition throughput is increased by 2.44 times.

With the development of knowledge graph in the field of artificial intelligence,there is an increasing demand to integrate knowledge graph from different sources to obtain a big knowledge graph with wider coverage.Ontology is the superstructure that can guide the construction of knowledge graph.To solve the problem of ontology alignment in knowledge graph fusion,this paper proposes an ontology alignment method based on self-attention model to combine multidimensional similarities.Firstly,two concepts from two ontologies are multi-dimensional measured by string-based,semantic-based and structure-based similarities.Then,self-attention model is used to combine above similarity calculations to judge whether the two concepts are similar or not and align them.Experiments on public datasets show that,compared with existing ontology alignment methods,the proposed method can obtain better alignment results by aggregating multi-dimensional similarity features.

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Computer Measurement & Control is the property of Magazine Agency of Computer Measurement & Control and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Aiming at the problem of the remaining useful life prediction accuracy being too low due to the complex operating conditions of the aviation turbofan engine data set and the original noise of the sensor, a residual useful life prediction method based on spatial–temporal similarity calculation is proposed. The first stage is adaptive sequence matching, which uses the constructed spatial–temporal trajectory sequence to match the sequence to find the optimal matching sample and calculate the similarity between the two spatial–temporal trajectory sequences. In the second stage, the weights of each part are assigned by the two weight allocation algorithms of the weight training module, and then the final similarity is calculated by the similarity calculation formula of the life prediction module, and the final predicted remaining useful life is determined according to the size of the similarity and the corresponding remaining life. Compared with a single model, the proposed method emphasizes the consistency of the test set and the training set, increases the similarity between samples by sequence matching with other spatial–temporal trajectories, and further calculates the final similarity and predicts the remaining use through the weight allocation module and the life prediction module. The experimental results show that compared with other methods, the root mean square error (RMSE) index and the remaining useful life health score (Score) index are reduced by 12.6% and 14.8%, respectively, on the FD004 dataset, and the RMSE index is similar to that in other datasets; the Score index is reduced by about 10%, which improves the prediction accuracy of the remaining useful life and can provide favorable support for the operation and maintenance decision of turbofan engines.

To solve the problem that existing algorithms were difficult to effectively identify sensitive sequences in genomic data for long-read with high error rate, a recognition algorithm using hybrid CPU and GPU parallel computing, called CGPU-F3SR, was proposed. Firstly, the long-read in genomic data were partitioned into multiple short-read, and the Bloom filtering mechanism was used to avoid repeated calculation of the short-read. Secondly, the k-mer coding strategy was used to extract in parallel the error information of all short-read, the recognition accuracy was promoted by improving the sequence similarity calculation model. Finally, CPU and GPU were used to coordinate and parallel to accelerate the calculation of short-read similarity to improve recognition efficiency. As a result, both two types of sensitive sequences including short tandem repeats and disease related sequences could be identified efficiently and accurately from genome data for long-read with high error rate. The experimental results of recognizing sensitive sequences from genomic data for long-read with length 100-400 kbp each show that, compared with existing parallel algorithm, the average recognition accuracy and precision rate of proposed CPU/GPU parallel algorithm CGPU-F3SR are increased by 7.77% and 43.07% respectively, its average false positive rate is reduced by 7.41%, and its average recognition throughput is increased by 2.44 times. [ABSTRACT FROM AUTHOR]

Unified Modeling Language (UML) models are main artifacts of software applications and valuable assets of software projects. The similarity calculation of UML models is an effective way to model reuse, model validation, model retrieval and pattern detection, which can help developers reduce development cost. Existing studies mainly focus on UML design models such as class diagrams and object diagrams. Yet, there is little research on calculating similarity for use case models despite their importance for software requirements reuse and validation. Thus, this paper proposes a similarity calculation approach for UML use case models. This approach first transforms the use case models into use case graphs and then adopts the deep learning technique — Similarity Graph Neural Network (SimGNN) to calculate the structural similarity between use case graphs. Then it calculates the semantic similarity of use case models with Term Frequency-Inverse Document Frequency (TF-IDF) and Cosine similarity. Finally it syntheses the results of semantic similarity and structural similarity of use case graphs. The effectiveness and efficiency of this approach was evaluated by comparing to two other approaches: (1) Cosine similarity+TF-IDF and (2) Cosine similarity+edit distance. The results showed that the our approach has better performance in terms of both effectiveness and efficiency. • A novel deep learning-based approach to calculate similarity of use case diagrams. • The integration of both semantic and structural information of use case diagrams. • A set of comparison experiments on a real dataset of use case diagrams. • Our approach is both effective and efficient. • Useful for model reuse, model validation, model retrieval and pattern detection. [ABSTRACT FROM AUTHOR]

• E-learning, as a part of social networking services, can effectively improve the effectiveness of teaching. • This article analyzes the effectiveness of interactive E-Learning mode based on artificial intelligence in English learning research. • We validated the effectiveness and feasibility of our model by demonstrating its ability to distinguish polysemy and its performance on word similarity datasets in experiments. E-learning, as a part of social networking services, can effectively improve the effectiveness of teaching. At the same time, in the process of knowledge dissemination, digital entertainment models can also enhance students' interest in learning. This article analyzes the effectiveness of interactive E-Learning mode based on artificial intelligence in English learning research. The semantics of English are relatively rich, and the same meaning can be expressed using different words or expressions. By calculating the similarity of sentences, students can understand the connections and differences between different expressions. Students can master the diversity and flexibility of word meanings by comparing different sentences, thereby improving their vocabulary and expression ability. By combining word vectors with semantic similarity, we can better measure the similarity between sentences, thereby improving the efficiency and accuracy of natural language processing. We validated the effectiveness and feasibility of our model by demonstrating its ability to distinguish polysemy and its performance on word similarity datasets in experiments. [ABSTRACT FROM AUTHOR]

Sudden natural disasters have the characteristics of complexity, unpredictability and frequency. To better manage and analyze sudden natural disasters promptly with the help of historical natural disaster cases, this study adopts the method of fused case structure to calculate the similarity between sudden disaster cases. Based on the disaster information of historical natural disaster cases, this paper aims to perform similarity measures for sudden natural disaster cases that contain textual information, data information and geographic location information at the same time. Taking storm surge disasters as examples, we designed a hierarchical case structure of "vertex-edge-label" based on the characteristics of sudden natural disaster cases. Then, we calculated the case similarity based on three aspects of sudden natural disasters, which were "case scenario", "disaster damage" and "case structure". Finally, we aggregated multiple factors to obtain the similarity between storm surge cases and conducted experiments on the historical storm surge cases in China. The result verified the feasibility and effectiveness of the method and showed a higher accuracy of the established aggregated multifactor method compared with the benchmark method. [ABSTRACT FROM AUTHOR]

In the information age, the amount of data is growing exponentially.However, different data sources are heterogeneous, which makes it inconvenient to share and multiplex data.With the rapid development of semantic network, ontology mapping is an effective method to solve this problem.The core of ontology mapping is ontology similarity calculation.Therefore, a calculation method based on graph convolution network was proposed.Firstly, ontologiesare modeled as a heterogeneous graph network, then the graph convolution network was used to learn the text embedding rules, which made ontologies were definedin global unified representation.Lastly, multisource data fusion was completed.The experimental results show that the accuracy of the proposed method is higher than other methods, and the accuracy of multi-source data fusion was effectively improved.

In the information age, the amount of data is growing exponentially. However, different data sources are heterogeneous, which makes it inconvenient to share and multiplex data. With the rapid development of semantic network, ontology mapping is an effective method to solve this problem. The core of ontology mapping is ontology similarity calculation. Therefore, a calculation method based on graph convolution network was proposed. Firstly, ontologiesare modeled as a heterogeneous graph network, then the graph convolution network was used to learn the text embedding rules, which made ontologies were definedin global unified representation. Lastly, multisource data fusion was completed. The experimental results show that the accuracy of the proposed method is higher than other methods, and the accuracy of multi-source data fusion was effectively improved.

A textual data processing task that involves the automatic extraction of relevant and salient keyphrases from a document that expresses all the important concepts of the document is called keyphrase extraction. Due to technological advancements, the amount of textual information on the Internet is rapidly increasing as a lot of textual information is processed online in various domains such as offices, news portals, or for research purposes. Given the exponential increase of news articles on the Internet, manually searching for similar news articles by reading the entire news content that matches the user’s interests has become a time-consuming and tedious task. Therefore, automatically finding similar news articles can be a significant task in text processing. In this context, keyphrase extraction algorithms can extract information from news articles. However, selecting the most appropriate algorithm is also a problem. Therefore, this study analyzes various supervised and unsupervised keyphrase extraction algorithms, namely KEA, KP-Miner, YAKE, MultipartiteRank, TopicRank, and TeKET, which are used to extract keyphrases from news articles. The extracted keyphrases are used to compute lexical and semantic similarity to find similar news articles. The lexical similarity is calculated using the Cosine and Jaccard similarity techniques. In addition, semantic similarity is calculated using a word embedding technique called Word2Vec in combination with the Cosine similarity measure. The experimental results show that the KP-Miner keyphrase extraction algorithm, together with the Cosine similarity calculation using Word2Vec (Cosine-Word2Vec), outperforms the other combinations of keyphrase extraction algorithms and similarity calculation techniques to find similar news articles. The similar articles identified using KPMiner and the Cosine similarity measure with Word2Vec appear to be relevant to a particular news article and thus show satisfactory performance with a Normalized Discounted Cumulative Gain (NDCG) value of 0.97. This study proposes a method for finding similar news articles that can be used in conjunction with other methods already in use.