利用置信度重取样的SemiBoost-CR分类模型

结合半监督学习和集成学习方法,提出了一种基于置信度重取样的SemiBoost-CR分类模型.给出了基于标注近邻与未标注近邻的置信度计算公式,按照置信度重采样,不仅选取一定比例置信度较高的未标注样本,而且选取一定比例置信度较低的未标注样本,分别以不同的策略加入到已标注的训练样本集,引入置信度高的未标注样本,用以提高基分类...     

半监督文本分类综述

文本分类是人们日常工作中经常遇到的问题,也是机器学习的重要研究内容.半监督学习算法同时考虑有标记和无标记数据,能显著提升学习效果.给出了文本分类的定义和特点,介绍了传统的监督学习分类算法和评价指标,对半监督文本分类的特点和基础理论进行了分析,并具体介绍了一些半监督文本分类算法,如贝叶斯方法和正则化方法.     

Heuristics for the multi-period orienteering problem with multiple time windows

We present the multi-period orienteering problem with multiple time windows (MuPOPTW), a new routing problem combining objective and constraints of the orienteering problem (OP) and team orienteering problem (TOP), constraints from standard vehicle routing problems, and original constraints from a real-world application. The problem itself comes from a real industrial case. Specific route duration constraints result in a route feasibility subproblem. We propose an exact algorithm for this subproblem, and we embed it in a variable neighborhood search method to solve the whole routing problem. We then provide experimental results for this method. We compare them to a commercial solver. We also adapt our method to standard benchmark OP and TOP instances, and provide comparative tables with state-of-the-art algorithms.     

网页驻留时间约束的用户访问序列挖掘

对Web用户的访问序列进行分析,可以发现用户的爱好、兴趣、习惯等因素,为Web网站的升级修正提供必要的信息支持,提出一种通过对用户访问序列进行分析的数据挖掘方法,该方法采用网页驻留时间为参数来约减会话序列中的网页数,压缩频繁访问序列的规模。实验结果表明,该算法可以降低挖掘成本,为Web用户的商业数据挖掘提供有益的借鉴。     

基于结构和纹理特征融合的场景图像分类

利用整体结构特征和局部纹理特征的优势,采用两级分类器对场景图像进行分类。第1级分类器利用全局结构信息得到候选类别,并通过分类结果判定相似类别对;第2级分类器则利用局部纹理信息区分相似类别,采用分类器的级联综合利用场景图像的整体结构信息和局部纹理信息。实验结果表明,该方法能够做到不同场景类别鲁棒分类,有效区分相似场景类别,提高场景图像的分类准确率。     

基于后验概率制导的B-KNN文本分类方法

针对K最近邻(KNN)方法分类准确率高但分类效率较低的特点,提出基于后验概率制导的贝叶斯K最近邻(B-KNN)方法。利用测试文本的后验概率信息对训练集多路静态搜索树进行剪枝,在被压缩的候选类型空间内查找样本的K个最近邻,从而在保证分类准确率的同时提高KNN方法的效率。实验结果表明,与KNN相比,B-KNN的性能有较大提升,更适用于具有较深层次类型空间的文本分类应用。     

A two-stage feature selection method for text categorization

Feature selection for text categorization is a well-studied problem and its goal is to improve the effectiveness of categorization, or the efficiency of computation, or both. The system of text categorization based on traditional term-matching is used to represent the vector space model as a document; however, it needs a high dimensional space to represent the document, and does not take into account the semantic relationship between terms, which leads to a poor categorization accuracy. The latent semantic indexing method can overcome this problem by using statistically derived conceptual indices to replace the individual terms. With the purpose of improving the accuracy and efficiency of categorization, in this paper we propose a two-stage feature selection method. Firstly, we apply a novel feature selection method to reduce the dimension of terms; and then we construct a new semantic space, between terms, based on the latent semantic indexing method. Through some applications involving the spam database categorization, we find that our two-stage feature selection method performs better.     

Least squares twin support vector machines for pattern classification

In this paper we formulate a least squares version of the recently proposed twin support vector machine (TSVM) for binary classification. This formulation leads to extremely simple and fast algorithm for generating binary classifiers based on two non-parallel hyperplanes. Here we attempt to solve two modified primal problems of TSVM, instead of two dual problems usually solved. We show that the solution of the two modified primal problems reduces to solving just two systems of linear equations as opposed to solving two quadratic programming problems along with two systems of linear equations in TSVM. Classification using nonlinear kernel also leads to systems of linear equations. Our experiments on publicly available datasets indicate that the proposed least squares TSVM has comparable classification accuracy to that of TSVM but with considerably lesser computational time. Since linear least squares TSVM can easily handle large datasets, we further went on to investigate its efficiency for text categorization applications. Computational results demonstrate the effectiveness of the proposed method over linear proximal SVM on all the text corpuses considered.     

Moving-in-pulse duration model-based target integration method for HSV-borne high-resolution radar

Hypersonic vehicles (HSVs) offer great advantages over airplanes and satellites, owing to their hypersonic speeds and flexible trajectories. However, HSV-borne radar may suffer from performance degradation if traditional target integration methods with stop-and-go (SAG) approximation are applied; this is especially true for high resolution radar and applications that utilize long-time coherent integration. In this paper, we consider the effects of moving-in-pulse duration (MPD) to derive an HSV-borne radar signal model without SAG approximation by specifically characterizing platform motion during the transmission and reception of a pulse. An explicit formula for wavenumber domain echo is derived using the 2-D joint stationary phase method. To mitigate the pulse-dependent echo distortion induced by the MPD effect, a target integration method based on the omega-K algorithm is proposed; this method employs revised filters for bulk focusing and revised Stolt interpolation for differential focusing to improve the overall focusing quality. The paper discusses the ambiguity functions of in-range MPD echo models, and describes the performance metrics for the integration results in both dimensions, including impulse response width (IRW), peak sidelobe ratio (PSLR), and integrated sidelobe ratio (ISLR). Simulation results have verified the effectiveness of the proposed method in MPD circumstances.     

新的文本分类特征选择方法研究

特征降维是文本分类过程中的一个重要环节。在现有特征选择方法的基础上,综合考虑特征词在正类和负类中的分布性质,综合四种衡量特征类别区分能力的指标,提出了一个新的特征选择方法,即综合比率(CR)方法。实验采用K-最近邻分类算法(KNN)来考查CR方法的有效性,实验结果表明该方法能够取得比现有特征选择方法更优的降维效果。