稀疏性

使用BP神经网络缓解协同过滤推荐算法的稀疏性问题

Employing BP Neural Networks to Alleviate the Sparsity Issue in Collaborative Filtering Recommendation Algorithms

利用合成孔径雷达（SAR）图像中目标的后向散射特性和目标散射中心的理论，分析了SAR图像数据稀疏性的成因。

The sparsity of SAR image data was discussed based on the backscattering properties of targets in the image .

一种保持对称性、稀疏性的拟Newton法

A kind of sparse symmetric quasi-Newton method

设计基于信息论和稀疏性的两种确定最佳分割尺度的算法，它们都能使分割后的CCD自然图像达到伪平衡状态。

Designed two algorithms to determine the optimal segmentation scale . One based on the cross-information , another based on sparse decomposition parameters .

改进方案在一定程度上解决了CF推荐算法遇到的矩阵稀疏性问题、冷启动问题和可扩展性差问题。

This improved program has solved the scalability , sparsity , cold start problems in a certain extent .

目前MRI重构算法只利用MRI图像稀疏性表示或只利用基于其局部光滑性的先验知识，重构效果不理想。

The current MRI reconstruction algorithms simply use either the sparse priors or the local smooth priors of MRI image , and result in the inferior reconstruction .

针对机械CAD三角网格模型中普遍存在的网格稀疏性和非均匀性分布等特点，提出了一种特征的二次提取及表面分割混合算法。

Aiming at the characteristics of sparseness and non-uniformity of meshes that is popular in triangle meshes of mechanical CAD model , a new approach to feature segmentation was proposed .

压缩传感（CS）理论是在已知信号具有稀疏性或可压缩性的条件下对信号进行采集、编解码的新理论。

Compressed Sensing ( CS ) theory is a novel data collection and coding theory under the condition that signal is sparse or compressible .

并在此基础上，在再生核Hilbert空间又提出一种稀疏性回归算法。

Also , a kind of sparse regression algorithm is presented based on the greedy algorithm in the reproducing kernel Hilbert space .

K空间数据具有变换稀疏性，可使用小波变换对K空间数据进行变换，得到具有稀疏性的数据空间，再进行稀疏采样成像。

K-space data has sparseness after transformation , in order to achieve sparse data space , we can use wavelet transform to transform the K-space data , then use the sparse sampling imaging .

仿真结果表明：固定尺度最小二乘支持向量机在训练各种样本数据集时，有效地避开了LS-SVM中的稀疏性问题，且训练速度快，同时具有良好的预测精度。

The simulation results indicate that fixed size LS_ - SVM shortens the training time enormously and possesses good predicting precision on different datasets .

该算法将CT图像的梯度稀疏性结合到ART图像重建中，在每次迭代中的投影操作结束后用梯度下降法调整全变差，减小图像梯度的1范数。

The method , which combines the gradient sparsity of CT images and ART , reduces the 1 norm of the image gradient by regulating the total variation with the gradient descend method after completing the projection on the corresponding hyperplane in each iteration .

贝叶斯方法可用来实现模型中超参数的自适应，同时保持SVR稀疏性和凸二次规划的优点。

Bayesian methods are used to implement model adaptation , while keeping the merits of support vector regression , such as sparseness and convex programming .

本文基于lk范数与稀疏性之间的关系，提出了一种基于lk范数的正则化方法，该方法在抑制信号噪声的同时，实现了信息的高精度稀疏表示。

Based on the relation between sparsity and lk norm , a regularization method is presented for the sparse representation of signals .

k-means聚类通常采用欧氏距离作为距离度量方法，但由于高维空间数据存在噪声且具有稀疏性，使得聚类效果显著降低，影响图像的表达。

The performance of k-means clustering severely degraded when Euclidean distance was used as the similarity measurement method because of the existence of the sparsity and noise in high-dimensional data .

同时，基于EEG信号的可稀疏性表示假设，提出了面向特征抽取的EEG信号结构自适应稀疏分解模型（SSDM），为本文的癫痫特征自动检测理论与算法奠定了模型基础。

Then , a structure adaptive sparse decomposition model ( SSDM ) is proposed for EEG signals , based on the hypothesis that EEG signals can be of sparse representation .

在机器学习方式的冗余字典构造中，提出了一种基于代理函数优化的稀疏性字典学习算法，并应用于Chirp回波冗余字典的构造。

In the construction of redundant dictionary with machine learning approach , a sparse dictionary learning algorithm based on optimization of surrogate function is put forward and applied to the construction of redundant dictionary for Chirp echoes .

该算法利用系统已经产生的Feed分类结果，计算类中各个项之间的相似性，然后再通过加权的方法对未评分项进行预测评分，从而有效降低了数据的稀疏性。

The algorithm uses the results of Feeds classification through system , to calculate the similarity between the various items , and then through the method of weighted , to forecast the score of items which did not been score , thus effectively reducing the sparsity of data .

传统的对RDF（ResourceDescriptionFramework）数据存储的方法主要采用基于关系数据库方式的三元组表，但由于RDF数据的稀疏性特点，使得这种存储方式的存储空间利用率和查询效率都不高。

Traditional RDF ( Resource Description Framework ) storage systems use relational database system to manage RDF data using a triple table , but due to the sparse characteristic of RDF data , those storage systems did not show efficient storage usage and high query performance .

协同过滤是目前电子商务推荐系统中广泛使用的、最成功的推荐算法，但还存在诸如稀疏性（sparsity）、冷启动（cold-start）、可扩展性（scalability）等制约其进一步发展的瓶颈问题。

Collaborative filtering is the most successful and widely used recommendation algorithm in E-commerce recommender systems currently . However , there exist some bottleneck problems in collaborative filtering , such as sparsity , cold-start and scalability .

算法用对称半正定矩阵作为元素目标函数的Hessian阵的近似，使得其和仍然保持目标函数的Hessian阵的某种稀疏性。

In the method , the symmetric positive semidefinite matrices are updated to approximate the Hessian matrices of the elemental objective functions and their sum to approximate the Hessian matrix of the objective function .

该方法利用了语音信号在DCT域的近似稀疏性，结合最优观测算法找出DCT基对应的最优观测矩阵，然后采用求得的最优观测矩阵对语音信号进行投影观测。

Knowing the approximate sparsity of speech signal in the DCT domain , the method first uses the optimized observation matrix algorithm to find the optimized observation matrix corresponding to the DCT matrix and then uses the acquired matrix to project the speech signals .

本文通过对当前B2C网站的电子商务个性化推荐系统分析，发现传统的推荐系统有如下问题：数据稀疏性问题，用户购买或评分的只占总商品数的1%左右；

There are several problems in traditional systems from the current B2C website electronic commerce personalization recommendation system : data sparsity , the commodities which are purchased or rated by users only occupy the total commodity number about 1 % ;

结合大规模电力系统修正方程组高维超稀疏性以及短向量的特点，提出以Krylov子空间方法研究电力系统方程计算问题。

This paper utilizes two characteristics of the large scale power system : the correction equation is large sparse and the vector is short , and studies the power system computation based on Krylov subspace method .

在上一章的基础上，第三章结合回声抵消器的特点，基于回声路径冲击响应稀疏性和精确块技术，提出了一种新的BEPNLMS自适应算法；

Chapter 3 proposes and studies a new adaptive Algorithm ( BEPNLMS ) based on the sparseness of the echo path and Block-Exact Techniques .

深入研究了基于原子分解的气体信号时频分析方法，对多种分解方法进行分析解释，提出了匹配追踪的Gram-Schmidt正交化和调整加权系数的改进方法，提高了分辨率和稀疏性，缩短了处理时间。

A detailed study of frequency analysis method of gas signal based on atomic decomposition is conducted . Improvement method of match pursuit based on the Gram-Schmidt orthogonalization and adjusting weighting coefficients is proposed to improve resolution and sparsity and reduce processing time .

DTN网络由于节点分布的稀疏性、节点移动的随机性、消息通信的不确定性使得网络拓扑频繁变化、通信链路经常中断、传输延迟相对较长，直接影响DTN网络的性能。

Due to the sparse distribution of the DTN nodes , the randomness of node mobility and the uncertainty of communication , the network topology frequently changes , communication links are often interrupted and the propagation delay is relatively long , which directly affect the DTN network performance .

这种新方法保持其修正矩阵的稀疏性。

The method retains the sparsity of secant updated matrix es .

稀疏性理论已被成功应用于许多机器学习方法中。

Sparsity has been successfully used to develop more efficient learning machines .

我们采用效率的观点，讨论了流网络的稀疏性问题。

Sparseness of flow networks is revisited from the viewpoint of efficiency .