Search

With the enormous growth of the content and users of social network, user portrait based on social network have been widely used in industrial applications such as artificial intelligence, recommendation systems, etc. The issue has been extensively studied, and most of them have focused on user attribute mining and behavior prediction. However, there is little research on the intrinsic relationship between big data-based interests and skills and occupational adaptability. To clarify this issue, in this article, we first collect a large amount of user data from LinkedIn. Then, we filter high-frequency interests and skills, and explore relevant characteristics of interests and skills through relevant analytical methods, found that there are many association rules. Next, the users are grouped according to the characteristics of occupational adaptability, and the relationship between the association rules of interests and skills and user adaptability is studied. Finally, designed an occupational adaptive classifier based on association rules. This article reveals the connections and dependencies between human interests and professional skills, as well as the impact of these association rules on user career adaptability, also discusses the prospects of these results in industrial applications. We hope that our research results will provide a solid theoretical foundation for other areas of research and industrial applications, such as career adaptability judgment, interest training, career development, career recommendation, etc.

© 2019 IEEE. The Internet improves the speed of information dissemination, and the scale of unstructured text data is expanding and increasingly being used for mass communication. Although these large amounts of data meet the infinite demand, it is difficult to find public focus in a timely manner. Therefore, information extraction from big data has become an important research issue, and there are many published studies on big data processing at home and abroad. In this paper, we propose a multi-feature keyword extraction method, and based on this, an artificial intelligence driven big data MFE scheme is designed, then an application example of the general scheme is expanded and detailed. Taking news as the carrier, this scheme is applied to the algorithm design of hot event detection. As a result, a multi-feature fusion clustering algorithm is proposed based on user attention with two main stages. In the first stage, a multi-feature fusion model is developed to evaluate keywords, and this model combines the term frequency and part of speech features. We use it to extract keywords for representing news and events. In the second stage, we perform clustering and detect hot events in accordance with the procedure, and during the composition of news clusters, we analyze several variadic parameters in order to explore the optimal effectiveness. Then, experiments on the news corpus are conducted, and the results show that the approach presented herein performs well.

With the development of mobile Internet, it is changing the way we communicate with others. Internet media have gradually become the main mobile crowdsourcing applications for information dissemination and user communication, including online news and social networks. However, the potential business opportunities have stimulated the emergence of a large number of spammers, who release false speech, advertisements, pornographic contents, and phishing websites on the media to gain commercial benefits, which seriously affects the experience of normal users. Therefore, in order to reduce the harm of false information, the research on the identification technology of spammers has been carried out extensively. However, the traditional technologies of identifying spammers involve high data costs and poor effects, and most of them are concentrated in the field of social networks, while less research is carried out in the field of online news. In this paper, we propose an effective technology of identifying online news comment spammers based on the label propagation algorithm (LPA), making full use of the user comment behaviors and contents. First of all, we collect a large amount of news and comments from NetEase News and label some users in the data as spammers or normal users manually to construct a labeled dataset. Then, a set of behavioral and semantic features are extracted and quantified from the user comment behaviors and comment contents by statistical analysis. Next, we propose the identification technology based on the LPA. Finally, the set of feature values is input into the proposed technology in different combinations, and experiments and evaluations are carried out to determine the most effective combination of features and improve the technology. The results show that the technology proposed in this paper involves a lower data cost but a better identification effect than some traditional technologies based on the supervised classifier.

Poverty eradication has been a challenge for humanity, especially when it comes to sustainability, if a family with a tendency to sink back into poverty can be detected, with necessary assistance provided in advance, such situation may be avoided. In order to detect them, we design a data-driven procedure to capture the relevance between poverty and related data, putting forward a deep neural multi-channel model to encode multi-type features, named Deep Poverty Forecast(DPF). In this paper, we also introduce a star graph to represent the relationship between family members and a data labeling method for supervised learning. Extensive experiments are conducted over five city datasets and our results show that our proposed framework has achieved much gain over previous methods. This solution applied at Data Platform of Poverty Reduction in Guangxi province has covered 84% of new families falling into poverty while the search space is reduced by over 90% according to the survey conducted by the local government. Our proposed framework can be easily applied to other family-related scenarios as well.

GPU not only is used in the field of graphic technology but also has been widely used in areas needing a large number of numerical calculations. In the energy industry, because of low carbon, high energy density, high duration and other characteristics, the development of nuclear energy cannot easily be replaced by other energy sources. Management of core fuel is one of the major areas of concern in a nuclear power plant, and it is directly related to the economic benefits and cost of nuclear power. The large-scale reactor core expansion equation is large and complicated, so the calculation of the diffusion equation is crucial in the core fuel management process. In this paper, we use CUDA programming technology on a GPU cluster to run the LU-SGS parallel iterative calculation against the background of the diffusion equation of the reactor. We divide one-dimensional and two-dimensional mesh into a plurality of domains, with each domain evenly distributed on the GPU blocks. A parallel collision scheme is put forward that defines the virtual boundary of the grid exchange information and data transmission by non-stop collision. Compared with the serial program, the experiment shows that GPU greatly improves the efficiency of program execution and verifies that GPU is playing a much more important role in the field of numerical calculations.

In recent years, researches on vocational skills are very extensive, most of them concentrated on vocational skills training, effect of vocational skill competition, vocational education and vocational skill identification and so on. But, there is very little research on the inner relationship among vocational skill based on big data. To address the issue, from LinkedIn we first collect tens of thousands of member profiles with vocational skills as empirical data. And then, we dig out high-frequency vocational skills and apply correlation analysis method to explore correlation characteristics of vocational skills. According to our studying, we figure out that there exist a large number of association rules, some of which have very high degrees of confidence, lift, and support. This research reveals the intrinsic association among different vocational skills. These relationships can reflect the general behavior and cognitive rules of human beings. We hope that our research results can provide a theoretical base for other research areas of vocational skills, such as vocational skill training, vocational skill mining, and identification of vocational skills.

In this paper we propose an approach that can find out the members for specific organizations based on user interaction on social networking Twitter. First we define three hypotheses to introduce user interaction on Twitter into organizational membership identification. Then we design the approach to mine colleagueship so as to recognize members for a specific organization. Given an organization with its public account and several known members as users on Twitter, our approach can dig out the users who affiliate the organization based on their interaction with the known members on Twitter. We also conducted an experiment to evaluate our approach. These experimental results show that our approach has a high recall rate. This research work tries to map the relations between users from virtual network to reality, which is very meaningful.

once an event occurs, usually there are a large number of online news to be released. How to quickly and accurately detect the hot events from the huge amount of online news is the focus and hotspot. Event detection and tracking technology is as a key technology to solve this problem. In this paper, we propose an approach to detect hot events from the online news stream in a timely manner and track the hot events. Based on the idea of single-pass clustering algorithm, this approach address the weight of keywords and proposes a new method to calculate similarity among news to track event. Through the analysis of the experimental results, we can find that this algorithm has a good effect on hot event detection.

"No comments, no news." Readers' online comments have become an important part of online news. Over the last few years, researches and applications on online news comments have been carried out. However, there are a few of researches about the distribution characteristics of online news comments. In this paper, first, we collect about one hundred thousand online news with their comments from six news websites in a given time period, and then study the online news comments'spatial distribution among online news and temporal distribution during their appearance. The result shows that the distribution of online news comments is greatly uneven. A small amount of news possess the vast majority of comments, and most of the news only have a few of online comments or even have no following comments. In addition, it shows the growth rate of online comments drops rapidly during a news conference. The findings can provide theoretical support to collection of online news comments, analysis of hot issues and the concern-specific behaviors characteristics of general public.

With the development of the mobile systems, we gain a lot of benefits and convenience by leveraging mobile devices; at the same time, the information gathered by smartphones, such as location and environment, is also valuable for business to provide more intelligent services for customers. More and more machine learning methods have been used in the field of mobile information systems to study user behavior and classify usage patterns, especially convolutional neural network. With the increasing of model training parameters and data scale, the traditional single machine training method cannot meet the requirements of time complexity in practical application scenarios. The current training framework often uses simple data parallel or model parallel method to speed up the training process, which is why heterogeneous computing resources have not been fully utilized. To solve these problems, our paper proposes a delay synchronization convolutional neural network parallel strategy, which leverages the heterogeneous system. The strategy is based on both synchronous parallel and asynchronous parallel approaches; the model training process can reduce the dependence on the heterogeneous architecture in the premise of ensuring the model convergence, so the convolution neural network framework is more adaptive to different heterogeneous system environments. The experimental results show that the proposed delay synchronization strategy can achieve at least three times the speedup compared to the traditional data parallelism.

Semantic web services (SWS) technology is developed to overcome the shortcomings of traditional standards, such as WSDL and UDDI, and enable maximal automation in all aspects of web service. But great improvement of capability in SWS-based service discovery is still desired. To address this issue, we present a distributed and semantic-matching-based approach for SWS publication and discovery by leveraging structured P2P technology. In this paper, first, we present our service matching rule which focuses on the inputs and outputs of SWS. Then, to realize this rule in open distributed environment and achieve the optimal match between a query and its qualified services, we propose not only a concept of ordered-concept-tree (OCT) to semantically sort the relevant concepts for service matching, but also a method to publish ontologies on structured P2P network to freely share and make full use of the concepts defined in different ontologies for OCT construction. Finally, we present a method for SWS publication on P2P network and an algorithm for service discovery. We also conduct experiments to evaluate our approach and the experimental results demonstrate its scalability, effectiveness, and callback capability to discover semantic web services.

Indicators of medical quality is the main basis of medical quality evaluation, predictive analyzing the indicators by using data mining technology, we can find and solve the problems of hospital quality management more timely. The study, basing on the medical homepage records, analyzes the key indicators of the hospital medical quality by decision tree C4.5 algorithm. Decision tree is a prediction model which can assist hospital to make related decision on medical quality management issues. By using decision tree C4.5 algorithm with appropriate modification, a high quality model of decision tree can be made. It can not only do the statistic of medical quality indicators, but also predict the indicators with high relative accuracy to improve hospital's medical quality management.

Semantic Web Service (SWS) technology is developed to overcome the shortcomings of traditional standards, such as WSDL and UDDI, and enable maximal automation in all aspects of Web Service provision and use. But great improvement of capability in SWS-based service discovery is still desired. To address this issue, we present a distributed approach for Semantic Web Service publication and discovery by leveraging structured P2P technology. In this paper, first, we introduce a semantics-based service matching method. Then, in order to apply the method to our approach, we propose several concepts and algorithms. Next, we present a method to publish SWS and implement it as a SWS registry. Finally, we design an approach to organize the nodes with the registry into a structured P2P network to cooperatively publish and discover SWS. We also conduct experiments to validate our approach and the results demonstrate its scalability and effectiveness.

Web ontologies, which usually appear as RDF graphs, are used to represent knowledge in web. In recent years, with the wide application of Semantic Web, large numbers of such web ontologies have appeared on the Internet, especially, in some virtual knowledge communities. These web ontologies are usually distributed in different sites and provide an amount of knowledge to query. But, it has become a pressing issue that, given a semantic query, how to efficiently gather the related knowledge from these web ontologies located in different sites to process it. To address this issue, in this paper, we propose a P2P-based approach to publish RDF triples in sharable web ontologies and freely sharing them in an open distributed environment. Given a query of SPARQL (Simple Protocol and RDF Query Language), this approach can automatically gather published RDF triples related to the query to process the query. As a knowledge sharing approach, our approach can directly share RDF triples coming from different web ontologies on different nodes. It overcomes limitations of those approaches which focus on how to locate related ontologies for a query. We also conducted three experiments to evaluate the effectiveness and the efficiency of our approach. The experimental results demonstrated that it is effective and efficient.

In order to efficiency teach and learn knowledge in academic courses for instructors and students, some systems are developed based on information and network technologies. However, these systems usually do not describe the knowledge in academic course formally. Students can not query course knowledge well and truly. Therefore, in this paper, we propose an approach based on web ontology and implement it to formally manage course knowledge and facilitate student querying it based on inference. In this approach, we focus on its architecture, implementation, and two special difficulties, i.e., how to aid instructors to create course ontology and how to convert a student's request into a structural query to reason course ontology.

In modern society, more and more people are suffering from some type of stress. Monitoring and timely detecting of stress level will be very valuable for the person to take counter measures. In this paper, we investigate the use of decision analytics methodologies to detect stress. We present a new feature selection method based on the principal component analysis (PCA), compare three feature selection methods, and evaluate five information fusion methods for stress detection. A driving stress data set created by the MIT Media lab is used to evaluate the relative performance of these methods. Our study show that the PCA can not only reduce the needed number of features from 22 to five, but also the number of sensors used from five to two and it only uses one type of sensor, thus increasing the application usability. The selected features can be used to quickly detect stress level with good accuracy (78.94%), if support vector machine fusion method is used.

OWL ontologies are used to represent knowledge in web. In recent years, with the wide application of Semantic Web, large numbers of OWL ontologies have appeared on Internet, especially, in some virtual knowledge communities. These ontologies are distributed in different sites and provide an amount of knowledge to query. But, it has become a pressing issue that, given a semantic query, how to efficiently gather the related knowledge from these ontologies located in different sites to process it. To address this issue, in this paper, we propose a P2P-based approach to publish axioms in sharable ontologies and freely sharing them in an open distributed environment. Given a query of SPARQL, this approach can automatically gather published axioms related to the query to process the query. As a knowledge sharing approach, our approach can directly share axioms coming from different ontologies on different nodes. It overcomes limitations of those approaches which focus on how to locate related ontologies for a query. We also conducted two experiments to evaluate the effectiveness and the efficiency of our approach. The experimental results demonstrated that it is effective and efficient.

In order to make recommendations to a user, a recommender mainly uses two approaches: content-basedfiltering approach and collaborative filtering approach. However, they both still have some shortcomings technically. The content-based approach is difficult to handle feature extraction as well as user intension prediction. The collaborative approach faces the hard issue of cold start problem and the matrix sparsity problem. In this paper, we present an novel hybrid recommendation approach based on Ontology and Neural Network in the movie domain. The approach combines content-based filtering and collaborativefiltering and a recommender can use them individually or use them both. The hybrid recommendation approach can tackle the traditional recommenders -- problems, such as feature extraction, intension prediction, matrix sparsity and cold start problems. Our experiments show that, our approach provides a good method to make recommendations to users.

Semantic Web Services (SWS) is proposed to ensure Web Service core missions automation, nevertheless, a dedicated Web Service discovery technology based on SWS is desired currently. For this reason, we propose an effective SWS matching algorithm to achieve service discovery based on semantics. In the paper, we first refine core properties and requirements to apply subsumption relation to involved SWS matching method, which is based on the semantic information of service's inputs and outputs. Then, we discuss a feasible service semantic matching algorithm in detail. Next, we implement and test it in related SWS registry. Finally, we point out its main properties. In fact, the algorithm proves to fully satisfy our expectations and be practical and feasible.

In previous work, we propose several mechanisms for distributed SWS publication and discovery. As same as some other applications of Semantic Web, they require users to provide concepts which are defined in domain ontology. However, users usually can not give an exact concept defined in ontology for a word in an open environment. For the reason, we propose a distributed ontology publication approach, through which we can publish ontology to P2P network, so that for a "word" user can find the ontologies that the "word" is defined in. Accordingly, a user can easily select an appropriate ontology to represent a concept for the "word". Then we apply the approach to our previous work to aid users to discover Web Service. With the testing work, the approach proves to be practical and effective approach to ontology publication and discovery.

In the paper, we analyze the shortages of Service- Oriented Analysis and Design (SOAD) methodologies, and then review the features of data flow diagram as a kind of requirement analysis technology. Finally, we propose an SOAD method using DFD diagram. Therefore, a process is provided as a guideline to aid analyst to accomplish high-quality service identification with the method. Through an illustration of the SOAD, we can conclude that the method has several strong points compared to methodologies else based on business process and enterprise architecture.