With the development of Internet,the amount of information on the network increases,and the content security issues on the network become more *** the study of web information filtering technology and a variety of imp...
详细信息
ISBN:
(纸本)9781467369541
With the development of Internet,the amount of information on the network increases,and the content security issues on the network become more *** the study of web information filtering technology and a variety of implementation methods,combining with UCL(uniform content locator) that is a new innovative technology concept,we discuss and design a new webpage filtering scheme,and propose a UCL-based trusted computing model of webpage content,which is using UCL technology to realize intelligent information filtering based on *** on the UCL semantic vector and Chinese corpus design,the model uses a reliability evaluation algorithm of webpage information content to realize reliability computing based on ***,the model combines with the integrity detection and the source authentication of information,thus forming a collaborative content filtering mechanism which is of high efficiency and can satisfy the filtering accuracy ***,we present the prototype *** have shown that due to less text content in UCL,the model is very suitable for fast processing and filtering of mass data,and has certain references and use values for network content security monitoring and information service mechanisms.
Deep learning,as a new unsupervised leaning algorithm,has strong capabilities to learn data *** work has shown that new features learned by deep learning algorithm help to improve the accuracy of cross-domain *** this...
详细信息
ISBN:
(纸本)9781467369541
Deep learning,as a new unsupervised leaning algorithm,has strong capabilities to learn data *** work has shown that new features learned by deep learning algorithm help to improve the accuracy of cross-domain *** this paper,we firstly propose a modified version of marginalized stacked denoising autoencoders(mSDA).We call it mSDA++ algorithm,which can learn excellent and low-dimensional features for training *** addition,we combine mSDA with EASYADAPT algorithm to further improve the accuracy of cross-domain *** we use SVM,mSDA,mSDA++,and EA+mSDA algorithms to do the cross-domain sentiment classification experiments on Amazon benchmark *** results show that EA+mSDA algorithm attains the best ***,the mSDA++ algorithm can accelerate the subsequent calculation and reduce the data storage space.
Foreign exchange rate forecasting catches many researchers interests in recent years. Problems of the foreign exchange rate forecasting model selection and the improvement on forecasting accuracy are not easy to be so...
详细信息
As a data integration technology, ontology has been widely used in the knowledge systems and domain knowledge representation. An ontology construction and fusion technology for large-scale data is very important. In t...
详细信息
As a data integration technology, ontology has been widely used in the knowledge systems and domain knowledge representation. An ontology construction and fusion technology for large-scale data is very important. In this paper, we propose a parallel ontology construction and fusion approach adapted to MapReduce framework based on the traditional ontology technology and MapReduce. The method separates the ontology model building process from the repeated calculation processes, and realizes the massive data integration. The evaluations tested on scientific literature data show the feasibility and efficiency of our approach.
In order to overcome the one-sidedness and limitations of a single subspace in feature extraction and classification, we propose a face recognition method that extracts features in double complementary space and utili...
详细信息
ISBN:
(纸本)9781467369541
In order to overcome the one-sidedness and limitations of a single subspace in feature extraction and classification, we propose a face recognition method that extracts features in double complementary space and utilize multi-decision for classification. In the feature extraction stage, we use ICA and LPE algorithm as the first layer in complementary space to extract global and local features of the face images. Then, for the purpose of solving the problem that features extracted by ICA are lack of classification information, we further extract classification information from the independent features extracted by ICA on the condition that the FLDA and DCV algorithm are used as the second layer in the complementary space. In the classification stage, the test samples are firstly projected to the independent subspace. After that, projected the samples which are difficult to identify to the LPE space and reclassified them. Finally, the results are been determined comprehensively. Experimental results on ORL database show that the proposed method can effectively improve the recognition rate.
Many machine learning methods have been applied on Named Entity Recognition (NER). Such methods generally build on a large manually-annotated training set. However, the training set is usually limited as human labelin...
详细信息
Many machine learning methods have been applied on Named Entity Recognition (NER). Such methods generally build on a large manually-annotated training set. However, the training set is usually limited as human labeling is costly and time consuming. Compare to the training set, the unlabeled corpus is usually much bigger and contains rich information about language. In this paper, a hybrid Deep Neural Network (DNN) is proposed to take advantage of the implicit information embedded in the un-labeled corpus. The experiments show that F1-score is improved from 85% to 90% (person name), from 75% to 81% (location name), and from 74% to 78% (organization name), compared with Conditional Random Fields (CRFs).
After studying the structure and model of existing navigation map data,this paper presented a design scheme of intelligent navigation data *** described the data model and data organization which was used by data plat...
详细信息
After studying the structure and model of existing navigation map data,this paper presented a design scheme of intelligent navigation data *** described the data model and data organization which was used by data platform,and designed conceptual model,logical model,physical model in *** paper achieved the construction of intelligent navigation data platform and provided strong support for various applications of telematics and intelligent tourism.
This paper put forwards a new single image restoration method which is self-adaptive to the dielectric layer *** method is used to solve the problem that existing single image visibility restoration method has limited...
详细信息
This paper put forwards a new single image restoration method which is self-adaptive to the dielectric layer *** method is used to solve the problem that existing single image visibility restoration method has limited adaptability for dielectric layer color in image acquisition ***,the method can remove influence caused by propagation medium layer from input image,realizing visibility restoration of the input *** show hazy image restoration method proposed by this paper has better adaption ability for dielectric layer color and recover an image.
Cyber-Physical systems (CPSs) represent a new generation of digital systems, where cyber entities and physical devices cooperate towards a set of common goals. The research presented in this paper aims to contribute t...
详细信息
Cyber-Physical systems (CPSs) represent a new generation of digital systems, where cyber entities and physical devices cooperate towards a set of common goals. The research presented in this paper aims to contribute to the development of CPSs by proposing an open architecture applicable in environmental monitoring, consisting of three layers. The paper provides a detailed description of each one of the CPS's main components. The bottom layer, composed of wireless sensor nodes, allows the acquisition of data and their transfer to the upper layers. The top and middle layers, composed of a measurements database, a knowledge base, a multi-agent society and web services, assure the proper operation of the CPS based on decision rules and complex data analyses. The proposed architecture provides a high degree of flexibility and scalability.
As a data integration technology,ontology has been widely used in the knowledge systems and domain knowledge *** ontology construction and fusion technology for large-scale data is very *** this paper,we propose a par...
详细信息
As a data integration technology,ontology has been widely used in the knowledge systems and domain knowledge *** ontology construction and fusion technology for large-scale data is very *** this paper,we propose a parallel ontology construction and fusion approach adapted to MapReduce framework based on the traditional ontology technology and *** method separates the ontology model building process from the repeated calculation processes,and realizes the massive data *** evaluations tested on scientific literature data show the feasibility and efficiency of our approach.
暂无评论