检索结果-内蒙古大学图书馆

2017 International Symposium on Computer Science and Artificial Intelligence（ISCSAI 2017）

作者： Li Xin Sun Guozi Li Huakang Jiangsu Key Lab of Big Data and Security and Intelligent Processing Nanjing University of Posts and Telecommunications Collaborative Innovation Center for Economics crime investigation and prevention technology State Key Laboratory of Mathematical Engineering and Advanced Computing

With the development of information technology, the application of big data in financial aspects becomes more and more deepening. However, in the aspect of bank loans, the accuracy of traditional user loan risk prediction models, such as KNN, Bayesian, are not benefit from the data *** paper proposes to use DNN algorithm to forecast the risk of user loan based on the difficulties of current overdue prediction and the excellent learning ability of DNN. This article uses user basic information, bank records, user browsing behavior, credit card billing records, and loan time information to evaluate whether users are delinquent. Firstly, this paper record bank records according to the transaction type, respectively, to generate income and spending data. Secondly, to sum the user browsing behavior also, and to record the average of credit card bill. In addition, in order to reduce the effect of eigenvalue size on the result, all characteristics are standardized. Finally, users who lack user information are discarded and the above fields are spliced. The spliced fields are the basic input for DNN. From the experimental results, DNN algorithm inc rease over 6% prediction than kNN, Bayes algorithm.

关键词： Overdue Prediction Bank Loans DNN

来源：评论

学校读者我要写书评

暂无评论

Deep hybrid residual learning with statistic priors for single image super-resolution

Deep hybrid residual learning with statistic priors for sing...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Risheng Liu Xiangyu Wang Xin Fan Haojie Li Zhongxuan Luo Key Laboratory for Ubiquitous Network and Service Software of Liaoning Province Department of Computing The Hong Kong Polytechnic University DUT-RU International School of Information & Software Engineering Dalian University of Technology

ISBN: (纸本)9781509060689

This paper considers single image super-resolution (SISR), which is an important low-level vision task and has various applications in multimedia society. Recently, deep neural networks have archived good performance on this field. But most of existing deep models are based on the fully data-dependent network architecture, thus missing majority of domain-knowledge of the super-resolution task. To address this limitation, we develop a new hybrid residual learning approach to leverage priors of SISR within the maximum a posteriori framework for network architecture design. We demonstrate that it can incorporate both image priors and data fidelity into the network, leading to a novel cascaded residual learning system for SISR process. Extensive experimental results on real-world images show that the proposed algorithm performs favorably against state-of-the-art methods.

关键词： Image resolution network architecture Radio frequency Training Degradation Dictionaries Convolution

来源：评论

学校读者我要写书评

暂无评论

Leveraging geometric correlation for input-adaptive facial landmark regression

Leveraging geometric correlation for input-adaptive facial l...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Yuyao Feng Risheng Liu Xin Fan Kang Huyan Zhongxuan Luo Key Laboratory for Ubiquitous Network and Service Software of Liaoning Province DUT-RU International School of Information & Software Engineering Dalian University of Technology Department of Computing The Hong Kong Polytechnic University

ISBN: (纸本)9781509060689

Facial analysis plays very important role in many vision applications, such as authentication and entertainments. The very early works in the 1990s mostly focus on estimating geometric deformations of facial landmarks to address this task. While in the past several years, more and more efforts have been made to directly learn an appearance regression for facial analysis. Though training regressions on controlled facial images can successfully capture the appearance variations, the performance of these appearance-based models are tightly related to the quantity and quality of the training data. In this paper, we develop a novel framework, named geometric correlated landmark regression (GCLR), to inherit the advantages but overcome limitations of these two categories of methods. Specifically, we first establish a landmark-to-landmark regression to estimate the geometry of facial images. By further incorporating a sparse coding term into the regression framework, we can successfully leverage the geometric correlations between the test image and the shape dictionary, thus significantly enhance the geometry regression performance. Experimental results on various challenging facial data sets verify the effectiveness and efficiency of GCLR.

关键词： Training Correlation Shape Face Geometry Databases Encoding

来源：评论

学校读者我要写书评

暂无评论

Least squares twin SVM decision tree for multi-class classification

Least squares twin SVM decision tree for multi-class classif...

引用

International Congress on Image and Signal Processing, BioMedical Engineering and Informatics

作者： Qing Yu Lihui Wang Tianjin Key Laboratory of Intelligence Computing and Network Security Tianjin University of Technology China

ISBN: (纸本)9781509037117

In order to solve multi-class classification problem in real world, we improved TSVM in this paper. We combined LSTSVM with partial binary tree to improve classification efficiency. Binary tree hierarchy can solved the inseparable regional issues in OVO-SVM and OVA-SVM classification. Experimental results show that it improved the classification accuracy. It also has better speed-up ratio than the OVO-SVM, OVA-SVM. It also reduced the training time. The time advantage is more obvious, especially the data set is large.

关键词： data mining the least squares support vector machine partial binary tree multi-class classification

来源：评论

学校读者我要写书评

暂无评论

Learning aggregated transmission propagation networks for haze removal and beyond

arXiv

引用

arXiv 2017年

作者： Liu, Risheng Fan, Xin Hou, Minjun Jiang, Zhiying Luo, Zhongxuan Zhang, Lei DUT-RU International School of Information Science and Engineering Key Laboratory for Ubiquitous Network Service Software of Liaoning Province Dalian University of Technology Dalian Department of Computing Hong Kong Polytechnic University

Single image dehazing is an important low-level vision task with many applications. Early researches have inves-tigated different kinds of visual priors to address this problem. However, they may fail when their assumptions are not valid on specific images. Recent deep networks also achieve relatively good performance in this task. But unfortunately, due to the disappreciation of rich physical rules in hazes, large amounts of data are required for their training. More importantly, they may still fail when there exist completely different haze distributions in testing images. By considering the collaborations of these two perspectives, this paper designs a novel residual architecture to aggregate both prior (i.e., domain knowledge) and data (i.e., haze distribution) information to propagate transmissions for scene ra-diance estimation. We further present a variational energy based perspective to investigate the intrinsic propagation behavior of our aggregated deep model. In this way, we actually bridge the gap between prior driven models and data driven networks and leverage advantages but avoid limitations of previous dehazing approaches. A lightweight learning framework is proposed to train our propagation network. Finally, by introducing a task-aware image separation formulation with a flexible optimization scheme, we extend the proposed model for more challenging vision tasks, such as underwater image enhancement and single image rain removal. Experiments on both synthetic and real-world images demonstrate the effectiveness and efficiency of the proposed framework. Copyright © 2017, The Authors. All rights reserved.

关键词： Domain Knowledge

来源：评论

学校读者我要写书评

暂无评论

Multiple influence maximization in social networks 10

Multiple influence maximization in social networks

引用

10th International Conference on Ubiquitous Information Management and Communication, IMCOM 2016

作者： Sun, Honglin Gao, Xiaofeng Chen, Guihai Gu, Jian Wang, Yongjian Shanghai Key Lab. of Scalable Computing and Systems Dept. of Computer Science and Engineering Shanghai Jiao Tong University Shanghai China Key Laboratory of Information Network Security of Ministry of Public Security Third Research Institute of Ministry of Public Security Shanghai China

ISBN: (纸本)9781450341424

Influence Maximization, a technique of social analysis to help marketers select a small group of users to promote their products with the goal of maximizing their influence spread, has been studied for a long time. However, previous researches did not consider the scenario that a company may intend to promote serval kinds of products in the same social network, where each user has different preferences for different categories. Considering this scenario in this paper, we propose the Multiple Influence Maximization (MIM) problem based on the assumption that one seed user can accept serval kinds of goods for free, and at the same time non-seed users have enough purchasing power to accept different promotions from their social friends. To address this issue, we design the corresponding greedy algorithm framework with an approximate radio of 1-1/e if the influence function is submodular. Besides, to simplify the computation of influence spread and considering the properties of real social influence propagation, we propose a novel influence quantitative computation model in which influence function fortunately is submodular. Applying this model into our MIM problem, the experiment results show that the greedy algorithm out-performs serval traditional heuristics in influence spread without loss of performance. © 2015 ACM.

关键词： Method of moments

来源：评论

学校读者我要写书评

暂无评论

Robust Multi-Exposure Image Fusion: A Structural Patch Decomposition Approach

Robust Multi-Exposure Image Fusion: A Structural Patch Decom...

引用

作者： Ma, Kede Li, Hui Yong, Hongwei Wang, Zhou Meng, Deyu Zhang, Lei Department of Electrical and Computer Engineering University of Waterloo WaterlooONN2L 3G1 Canada Department of Computing Hong Kong Polytechnic University Hong Kong Hong Kong Department of Computing. Hong Kong Polytechnic University Hong Kong Hong Kong Institute for Information and System Sciences and Ministry of Education Key Laboratory of Intelligent Networks and Network Security Xi'an Jiaotong University Xi'an710049 China Inst. for Info. and Syst. Sci. and Min. of Educ. Key Lab. of Intelligent Networks and Netwk. Secty. Xi'an Jiaotong University Xi'an710049 China

We propose a simple yet effective structural patch decomposition approach for multi-exposure image fusion (MEF) that is robust to ghosting effect. We decompose an image patch into three conceptually independent components: signal strength, signal structure, and mean intensity. Upon fusing these three components separately, we reconstruct a desired patch and place it back into the fused image. This novel patch decomposition approach benefits MEF in many aspects. First, as opposed to most pixel-wise MEF methods, the proposed algorithm does not require post-processing steps to improve visual quality or to reduce spatial artifacts. Second, it handles RGB color channels jointly, and thus produces fused images with more vivid color appearance. Third and most importantly, the direction of the signal structure component in the patch vector space provides ideal information for ghost removal. It allows us to reliably and efficiently reject inconsistent object motions with respect to a chosen reference image without performing computationally expensive motion estimation. We compare the proposed algorithm with 12 MEF methods on 21 static scenes and 12 deghosting schemes on 19 dynamic scenes (with camera and object motion). Extensive experimental results demonstrate that the proposed algorithm not only outperforms previous MEF algorithms on static scenes but also consistently produces high quality fused images with little ghosting artifacts for dynamic scenes. Moreover, it maintains a lower computational cost compared with the state-of-The-Art deghosting schemes.11The MATLAB code of the proposed algorithm will be made available online. Preliminary results of Section III-A [1] were presented at the IEEE International Conference on Image Processing, Canada, 2015. © 1992-2012 IEEE.

关键词： Image fusion

来源：评论

学校读者我要写书评

暂无评论

Improved Canopy-FCM algorithm based on MapReduce

Improved Canopy-FCM algorithm based on MapReduce

引用

International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)

作者： Qing Yu Zhimin Ding Tianjin Key Laboratory of Intelligence Computing and Network Security Tianjin University of Technology Tianjin China

ISBN: (纸本)9781509037117

In order to solve the problem that FCM algorithm is sensitive to initial clustering center, we use Canopy algorithm to conduct the quick rough clustering. In the meantime, to avoid the blindness of Canopy algorithm, we put forward an improved Canopy-FCM algorithm based on max-min principle. In allusion to the problem that FCM algorithm has high time complexity, this article use the parallel computing frame of MapReduce to design and realize the improved Canopy-FCM algorithm. Experimental result shows: improved Canopy-FCM algorithm based on MapReduce has better clustering quality and running speed than the Canopy-FCM and FCM algorithm based on MapReduce, and the improved Canopy-FCM algorithm based on Hadoop has better speed-up ratio than Canopy-FCM based on Standalone mode.

关键词： Clustering algorithms Algorithm design and analysis Signal processing algorithms Parallel processing Classification algorithms Blindness Data collection

来源：评论

学校读者我要写书评

暂无评论

ROSF: Leveraging Information Retrieval and Supervised Learning for Recommending Code Snippets

arXiv

引用

arXiv 2017年

作者： Jiang, He Nie, Liming Sun, Zeyi Ren, Zhilei Kong, Weiqiang Zhang, Tao Luo, Xiapu School of Software Dalian University of Technology Dalian China Key Laboratory for Ubiquitous Network and Service Software of Liaoning Province Dalian China State Key Laboratory of Software Engineering Wuhan University Wuhan China Key Laboratory for Ubiquitous Network Service Software of Liaoning Province Dalian China School of Computer Science & Technology Nanjing University of Posts and Telecommunications Nanjing China Department of Computing Hong Kong Polytechnic University Hong Kong Hong Kong

—when implementing unfamiliar programming tasks, developers commonly search code examples and learn usage patterns of APIs from the code examples or reuse them by copy-pasting and modifying. For providing high-quality code examples, previous studies present several methods to recommend code snippets mainly based on information retrieval. In this paper, to provide better recommendation results, we propose ROSF, Recommending cOde Snippets with multi-aspect Features, a novel method combining both information retrieval and supervised learning. In our method, we recommend Top-K code snippets for a given free-form query based on two stages, i.e., coarse-grained searching and fine-grained re-ranking. First, we generate a code snippet candidate set by searching a code snippet corpus using an information retrieval method. Second, we predict probability values of the code snippets for different relevance scores in the candidate set by the learned prediction model from a training set, re-rank these candidate code snippets according to the probability values, and recommend the final results to developers. We conduct several experiments to evaluate our method in a large-scale corpus containing 921,713 real-world code snippets. The results show that ROSF is an effective method for code snippets recommendation and outperforms the-state-of-the-art methods by 20% - 41% in Precision and 13% - 33% in NDCG. Copyright © 2017, The Authors. All rights reserved.

关键词： Supervised learning

来源：评论

学校读者我要写书评

暂无评论

A Performance Evaluation of Classification Algorithms for Big Data

引用

Procedia Computer Science 2017年 122卷 1100-1107页

作者： Mo Hai You Zhang Yuejin Zhang School of Information Central University of Finance and Economics Beijing China Network and Data Security Key Laboratory of Sichuan Province University of Electronic Science and Technology of China Chengdu China School of Information Systems &Management Heinz College Carnegie Mellon University Pittsburgh PA 1521 USA

The performance of two typical classification algorithms in Spark: random forest and naïve bayes are evaluated by using four metrics: classification accuracy, speedup, scaleup and sizeup. Experiments are performed on dataset and clusters of different scale. The results show that: (1) the accuracy of the two algorithms is high; (2) the increase of speedup is not linear. For the dataset with different size, the numbers of nodes is different when the speedup is the maximal; (3) the scaleup of random forest reaches its peak when the number of nodes is 2, and after that the scaleup decreases with the increase of the number of nodes;(4) for random forest, when the number of nodes is 2, the sizeup increases sharply with the increase of the size of dataset, and when the number of nodes is greater than 2, the sizeup increases more slowly with the increase of the size of dataset; for naïve bayes, when the number of nodes is smaller than 6, the sizeup increases with the increase of the size of dataset, when number of nodes is 6 and the size of dataset is larger than that of Sogou_5, the change of the sizeup is not obvious with the increase of the size of dataset.

关键词： Spark Random forest Naïve bayes Classification accuracy Speedup Scaleup Sizeup

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：