支持向量机算法在MOOC 课程答疑系统中的研究

随着互联网技术和近期MOOC课程的发展,智能答疑系统也受到了更多的关注,应用它能够及时给学生提供学生疑惑的问题答案。智能答疑系统通常包括问句理解、信息检索、答案抽取和选择三个主要部分,且问句分类是问句理解的关键,因为它的准确性将直接影响到最后答案的准确性。以高校计算机基础课程为实际背景,在已有基于支持向量机算法基础上,对该方法进行了改进,并通过训练集和测试集进行了验证。从实验结果看,该方法在高校计算机基础智能答疑系统中有比较好的应用效果。     

基于语义分类的比较句识别与比较要素抽取研究

比较是人们常用的评估不同事物优劣、异同的表达方式,利用机器识别比较句并进一步抽取比较要素是语言信息处理领域一项新颖又有实用价值的课题。该文依据比较句与比较要素之间是一种“你中有我,我中有你”的共生关系,将比较句识别与比较要素抽取两个任务合二为一完成;根据词意分类,构建由领域词典、情感词典、标记词典、普通词典构成的词典系统;根据汉语比较句句义分类,构建比较句识别与比较要素抽取规则库。以第四届中文倾向性评测(COAE2012)发布的测试语料为实验对象,该系统取得了较好的实验(评测)结果。     

单幅图像多尺度小波深度提取算法

针对浅景深图像中平滑前景区域深度提取误差大的问题,基于像素点分类思想对深度值进行修正,提出一种基于多尺度小波线索的、可同时面向单幅浅景深图像和广角图像的深度图提取算法.首先使用小波分析法在多个尺度下提取图像深度信息;然后提出自适应分类法并根据尺度与深度变化规律对像素点做深度修正,得到深度图;最后结合区域生长与边缘分割算法对深度图进行区域优化.为了加快深度计算,还提出了快速zerocount法以及多尺度加速法来满足标清视频实时处理要求.实验结果证明,采用文中算法获得的深度图相对深度正确,前景和背景区域深度一致性好.     

稀疏表示分类中遮挡字典构造方法的改进

针对稀疏表示分类算法中遮挡字典维数高且无冗余的问题,提出一种遮挡字典构造方法.首先通过图像分块得到各级的遮挡基图像;然后将所有互不相同的遮挡基图像按字典顺序转化为向量,并用这些向量作为遮挡字典的列,从而构造出维数相对较低且具有一定冗余度的遮挡字典.实验结果表明,该方法不仅明显提高了稀疏表示分类算法对遮挡人脸的识别率,而且还能通过减少图像的分块级数降低稀疏分解的耗时量,提高运算效率.     

基于多重分形特征的防空雷达目标分类方法

在介绍常规雷达飞机目标回波数学模型的基础上,利用多重分形测度分析手段,分析了常规雷达飞机目标回波的多重分形特性及其多重分形特征的提取方法,并从模式分类的角度,提出了利用多重分形特征对喷气式飞机、螺旋桨飞机和直升机等三类飞机目标进行分类的方法.实验表明,上述三类飞机的目标回波数据具有显著不同的多重分形特征,所定义的多重分形特征参数可以作为飞机目标分类和识别的有效特征,仿真实验验证了所提方法的有效性.     

面向情感特征的风景图像分类研究及其实现

视觉活动是认知外部世界的主要途径,视觉感知源于物体的外形、色彩及纹理。而颜色特征的不同对人类情感变化起着至关重要的作用。鉴于不同风景图像会产生不同的颜色特征和情感特点,如何建立图像颜色特征与用户评价之间的关系并加以鉴别具有重要的研究意义。该文在学习前人提出的各种颜色提取分类算法的基础上,结合图像的颜色直方图对风景图像的颜色特征进行提取,并利用支持向量机的方法对图像进行分类。实验表明,该方法取得了较好的准确率。     

电子商务模式的多维分类体系研究

电子商务模式的开展是实现电子商务应用的重要问题之一.为充分探讨电子商务模式的体系建立,在总结已有成果基础上,本文基于电子商务模式的分类讨论十大类基本的电子商务模式,分析构建多维电子商务模式分类体系.     

Processing of an East Mediterranean phosphate ore sample by an integrated attrition scrubbing/classification scheme (part one)

ABSTRACT

In this study, an Egyptian siliceous/calcareous sedimentary phosphorite was subjected to beneficiation by unsophisticated economic means. The mineralogical and chemical examination showed that the ore sample is a typical example for the East Mediterranean phosphorites with 21.80% P₂O₅, 10.93% acid insoluble and 16.61% loss on ignition. The primary crushed sample was classified into three size fractions without any milling and they were separately subjected to an integrated attrition scrubbing study. Screening and classification on the 2.30 mm and 0.071 mm sieves were sequentially introduced with the attrition process. The effect of solid%, attrition speed and time were studied and optimized for each size fraction. Results showed that three products were produced and could be categorized according to their P₂O₅ content into the coarse tailing (+2.30 mm), the attrition concentrate (?2.30 + 0.071 mm) and the phosphate fines (?0.071 mm). The P₂O₅ contents of these products reached 12.24%, 28.72% and 18.77%, with P₂O₅ distribution reached 6.74%, 51.75% and 41.51%, respectively. A global P₂O₅ of 28.72% was obtained with a recovery of about 52%. A tentative flow-sheet for the whole process was postulated.     

Gasoline quality prediction using gas chromatography and FTIR spectroscopy: An artificial intelligence approach

This paper reports on analysis of 45 gasoline samples with different qualities, namely, octane number and chemical composition. Measurements of data from gas chromatography and IR (FTIR) spectroscopy are used to gasoline quality prediction and classification. The data were processed using principal component analysis (PCA) and fuzzy C means (FCM) algorithm. The data were then analyzed following the neural network paradigms, hybrid neural network and support vector machines (SVM) classifier. The IR spectra were compressed and de-noised by the discrete wavelet analysis. Using the hybrid neural network and multi linear regression method (MLRM), excellent correlation between chemical composition of the gasoline samples and predicted value of the octane number was obtained. About 100% correct classification for six different categories of the gasoline was achieved, each of which has different qualities.     

Classification of vegetable oils by high-resolution <Superscript>13</Superscript>C NMR spectroscopy using chromatographically obtained oil fractions

¹³C NMR spectra of oil fractions obtained chromatographically from 109 vegetable oils were obtained and analyzed to evaluate the potential use of those fractions in the classification of vegetable oils and to compare the results with the NMR analysis of complete oils. The oils included the following: virgin olive oils from different cultivars and regions of Europe and north Africa; “lampante” olive, refined olive, refined olive pomace, hazelnut, rapeseed, high-oleic sunflower, corn, grapeseed, soybean, and sunflower oils; and mixtures of virgin olive oils from different geographical origins. Oils were divided into two sets of samples. The training set (98 samples) was employed to select the variables that resulted in significant discrimination among the different oil classes. By using stepwise discriminant analysis, more than 98% of correct validated assignments were obtained; these results were confirmed when applied to the test set (11 blind samples). Results suggest that the use of oil fractions considerably increases the discriminating power of NMR in the analysis of vegetable oils.