Robust support vector machine-trained fuzzy system

Because the SVM (support vector machine) classifies data with the widest symmetric margin to decrease the probability of the test error, modern fuzzy systems use SVM to tune the parameters of fuzzy if–then rules. But, solving the SVM model is time-consuming. To overcome this disadvantage, we propose a rapid method to solve the robust SVM model and use it to tune the parameters of fuzzy if–then rules. The robust SVM is an extension of SVM for interval-valued data classification.We compare our proposed method with SVM, robust SVM, ISVM-FC (incremental support vector machine-trained fuzzy classifier), BSVM-FC (batch support vector machine-trained fuzzy classifier), SOTFN-SV (a self-organizing TS-type fuzzy network with support vector learning) and SCLSE (a TS-type fuzzy system with subtractive clustering for antecedent parameter tuning and LSE for consequent parameter tuning) by using some real datasets. According to experimental results, the use of proposed approach leads to very low training and testing time with good misclassification rate.     

基于微量元素支持向量机在鼻咽癌模型预测中的应用*

背景：支持向量机目前已经在文本分类、手写识别、图像分类、生物信息学等诸多领域被成功应用。 目的：采用智能算法，将支持向量机算法与微量元素数据结合对鼻咽癌患者建模，以提高鼻咽癌识别正确率。 方法：基于微量元素数据，利用支持向量机对鼻咽癌患者、正常人、其他疾病患者样本建立分类模型。样品取自观察对象未染发头枕部紧贴头皮3 cm的头发。对样本进行的临床微量元素检测项目为6种元素锌、铜、铁、锰、镉、镍，加上年龄和性别共8项。采用高斯径向基函数为核函数、调节核函数参数C及σ以建立最佳支持向量机模型。 结果与结论：采用十折交叉验证法得到模型的识别率分别为81.71%和66.47%。结果表明，基于微量元素的支持向量机法建立的鼻咽癌分类模型能较好的把鼻咽癌样本从正常人、各种疾病患者样本中区分出来。     

基于波形特征和SVM的心电信号自动分类方法研究

目的:提出一种新的基于波形特征和SVM的心电信号自动分类实现方法。方法:定义并提取了基于时域特征、小波域特征和高阶统计量特征等三大类心电特征参数,将一次性直接求解多类模式的SVM方法应用于心电信号分类。结果:通过对心电数据库典型心律失常信号的分类测试,验证了所提出心电信号分类方法的有效性。结论:本方法的实现可以有效提高了分类识别精度和速度。     

Monitoring Dynamic Change of Land Cover Based on SVM and Satellite Images in Hanoi,Vietnam

An integral method, combining support vector machine （SVM） with remote-sensing analysis techniques, was explored to monitor Hnoi＇s dynamic change of land cover. The landsat thematic mapper （TM） image in 1993, the enhanced thematic mapper plus （ETM＋） image in 2000, and the image with the charge-coupled device camera （CCD） on the China-Brazil earth resources satellite （CBERS） in 2008 were used. Six land-cover types, including built-up areas, woodland, cropland, sand, water body and unused land, were identified. The detected results showed visually the rapid urban expansion as well as land-cover change of Hanoi from 1993 to 2008. There were 12 637.54 hm2 cropland decreased between 1993 and 2000, and 8 227.6 hm2 cropland decreased between 2000 and 2008. Compared with cropland, woodland firstly decreased and then increased, and the other types did not change significantly. The results indicate that CBERS dataset has the application potential in world resources researches.     

Prediction of Neural Tube Defect Using Support Vector Machine

Objective To predict neural tube birth defect （NTD） using support vector machine （SVM）. Method The dataset in the pilot area was divided into non overlaid training set and testing set. SVM was trained using the training set and the trained SVM was then used to predict the classification of NTD. Result NTD rate was predicted at village level in the pilot area. The accuracy of the prediction was 71.50% for the training dataset and 68.57% for the test dataset respectively. Conclusion Results from this study have shown that SVM is applicable to the prediction of NTD.     

Classification of Error-Related Negativity (ERN) and Positivity (Pe) potentials using kNN and Support Vector Machines

Error processing in subjects performing actions has been associated with the Event-Related Potential (ERP) components called Error-Related Negativity (ERN) and Error Positivity (Pe). In this paper, features based on statistical measures of the sample of averaged ERP recordings are used for classifying correct from incorrect actions. Three feature selection techniques were used and compared. Classification was done by means of a kNN and a Support Vector Machines (SVM) classifier. The use of a leave-one-out approach in the feature selection provided sensitivity and specificity values concurrently higher than or equal to 87.5%, for both classifiers. The classification results were significantly better for the time window that included only the ERN, as compared to time windows including also Pe.     

Brain volumetry: an active contour model-based segmentation followed by SVM-based classification

In this paper a novel automatic approach to identify brain structures in magnetic resonance imaging (MRI) is presented for volumetric measurements. The method is based on the idea of active contour models and support vector machine (SVM) classifiers. The main contributions of the presented method are effective modifications on brain images for active contour model and extracting simple and beneficial features for the SVM classifier. The segmentation process starts with a new generation of active contour models, i.e., vector field convolution (VFC) on modified brain images. VFC results are brain images with the least non-brain regions which are passed on to the SVM classification. The SVM features are selected according to the structure of brain tissues, gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF). SVM classifiers are trained for each brain tissue based on the set of extracted features. Although selected features are very simple, they are both sufficient and tissue separately effective. Our method validation is done using the gold standard brain MRI data set. Comparison of the results with the existing algorithms is a good indication of our approach's success.     

Spatial regularization of SVM for the detection of diffusion alterations associated with stroke outcome

In this paper, we propose a new method to detect differences at the group level in brain images based on spatially regularized support vector machines (SVM). We propose to spatially regularize the SVM using a graph Laplacian. This provides a flexible approach to model different types of proximity between voxels. We propose a proximity graph which accounts for tissue types. An efficient computation of the Gram matrix is provided. Then, significant differences between two populations are detected using statistical tests on the outputs of the SVM. The method was first tested on synthetic examples. It was then applied to 72 stroke patients to detect brain areas associated with motor outcome at 90 days, based on diffusion-weighted images acquired at the acute stage (median delay one day). The proposed method showed that poor motor outcome is associated to changes in the corticospinal bundle and white matter tracts originating from the premotor cortex. Standard mass univariate analyses failed to detect any difference on the same population.     

SVM算法用于黄连解毒汤3种指标性成分配伍浓度与药效作用相关性的研究

目的 将支持向量机算法用于中药复方物质基础研究,探讨建立能反映黄连解毒汤中指标性成分质量浓度与药效作用相关性的数学模型.方法 根据数学组合原理,将黄连解毒汤君(黄连)、臣(黄芩)、佐(黄柏)、使(栀子)4味药组合为11组,每组分别采用膜分离、大孔吸附树脂进行分离,得到22个部位.选用鼠性肾上腺髓质嗜铬细胞瘤PC-12细胞活性跟踪筛选,MTT法测定黄连解毒汤及各分离部位对过氧化氢(H2O2)、氯化钾(KCl)、连二亚硫酸钠(Na2S2O4)诱导的PC-12细胞损伤的影响.应用支持向量机(Support Vector Machine)算法考察标识黄连解毒汤的3种指标性成分小檗碱(以盐酸小檗碱计)、黄芩苷、栀子苷量与保护率之间的相关性,寻找3种成分浓度的优化配比关系.结果 小檗碱-黄芩苷-栀子苷的浓度比为1∶13∶12时,样品对过氧化氧致细胞损伤的保护率最高;小檗碱-黄芩苷-栀子苷的浓度比为13∶1∶3时,样品对连二亚硫酸钠造模的保护率最高;而对于氯化钾造成的损伤,未能得出最优浓度.结论 不同造模机制下得到的黄连解毒汤中指标性成分的最佳浓度比不同,该汤剂中指标成分适宜的浓度范围及比例需要进一步药理实验研究确定.