基于支持向量机的ICU中急性低血压预测模型研究

在重症监护室(ICU)的监护中,急性低血压(AHE)的发生严重威胁着患者的生命安全,临床上主要依靠医生的经验处置。本文运用医学信息学的理论,研究一种ICU中AHE发生的预测模型。利用ICU监护中血压变化的连续记录数据,分析发生与未发生AHE两者间平均动脉压(MAP)信号的变化趋势与特点,基于统计学习理论的支持向量机(SVM)方法,选取中位数、平均值等统计特征参数用于学习和训练,建立分类预测模型。在此基础上,对不同核函数构成的分类器和预测算法进行了比较分析。实验验证,本方法能够达到比较好的分类预测效果,有利于AHE发生的提前预测。     

数据挖掘与乳腺癌诊断的研究进展

数据挖掘是应用一系列技术从大型数据库中提取人们感兴趣的信息和知识,这些知识或信息是隐含的、事先未知而潜在有用的,可表示为概念、规则、规律、模式等形式。本文详尽介绍了数据挖掘技术产生的背景、概念,综述了近年数据挖掘技术在医学中以及辅助诊断乳腺癌的应用情况,并探讨了其在辅助诊断乳腺癌的应用前景、意义以及目前存在的问题。     

基于支持向量机的乳腺癌预后状态预测和疗效评估

乳腺癌是危害妇女健康的主要恶性肿瘤.目前基因与疾病关系的研究取得了一系列的成果,使得利用乳腺癌患者的基因信息来预测预后状态和评估治疗效果成为了可能.支持向量机(support vector machine,SVM)分类方法在实际二类分类问题的应用中显示出良好的学习和泛化能力,已被广泛地应用于诸多研究领域.本文采用支持向量机SVM、K-近邻法(K-nearest neighbor,K-NN)、概率神经网络(probabilistic neural network,PNN)、决策树(decision tree,DT)分类器,结合乳腺癌患者基因数据来预测患者的预后状态和评估治疗效果.结果表明:当使用高斯径向基核函数时,SVM通过5次交叉验证的最佳平均分类准确率达到了88.44%,优于K-NN(81.69%)、PNN(80.68%)和DT(71.19%)等分类器,表明该方法有望成为一种有效、实用的乳腺癌预后状态预测和治疗效果客观评价的工具.     

Diagnosis of osteoporosis by extraction of trabecular features from hip radiographs using support vector machine: An investigation panorama with DXA

Background

Lifespan and its quality can be improved by early diagnosis of osteoporosis. Analysis of trabecular boundness on digital hip radiographs could be useful for identifying subjects with low bone mineral density (BMD) or osteoporosis. The main aim of our study was to evaluate the ability of a kernel-based support vector machine (SVM) with respect to diagnosis and add to knowledge about the trabecular features of digital hip radiographs for identifying subjects with low BMD.

Method

In this paper we present an SVM kernel classifier-based computer-aided diagnosis (CAD) system for osteoporotic risk detection using digital hip radiographs. Initially, the original radiograph was intensified, then trabecular features such as boundness, orientation, solidity of spur and delta were evaluated and radial bias function (RBF) based discrimination was manifested. The next step was the evaluation of the diagnostic capability of the proposed method in order to spot subjects with low BMD at the femoral neck in 50 (50.7±14.3 years) South Indian women with no previous history of osteoporotic fracture. Out of 50 subjects, 28 were used to train the classifier and the other 22 were used for testing.

Results

The proposed system has achieved the highest classification accuracy documented so far by means of a fivefold cross-validation analysis with mean accuracy of 90% (95% confidence interval (CI): 82 to 98%); sensitivity and positive predictive value (PPV) were 90% (95% CI: 82 to 98%) and 89% (95% CI: 81 to 97%), respectively. Pearson's correlation was observed at the level of p<0.001, between extracted image trabecular features with age and BMDs measured by dual energy x-ray absorptiometry (DXA). Extracted image features also demonstrated significant differences between high and low BMD groups at the level of p<0.001.

Conclusion

Our findings suggest that the proposed CAD system with SVM would be useful for spotting women vulnerable to osteoporotic risk.     

Automatic detection of unstained viable cells in bright field images using a support vector machine with an improved training procedure

Detection of unstained viable cells in bright field images is an inherently difficult task due to the immense variability of cell appearance. Traditionally, it has required human observers. However, in high-throughput robotic systems, an automatic procedure is essential. In this paper, we formulate viable cell detection as a supervised, binary pattern recognition problem and show that a support vector machine (SVM) with an improved training algorithm provides highly effective cell identification. In the case of cell detection, the binary classification problem generates two classes, one of which is much larger than the other. In addition, the total number of samples is extremely large. This combination represents a difficult problem for SVMs. We solved this problem with an iterative training procedure ("Compensatory Iterative Sample Selection", CISS). This procedure, which was systematically studied under various class size ratios and overlap conditions, was found to outperform several commonly used methods, primarily owing to its ability to choose the most representative samples for the decision boundary. Its speed and accuracy are sufficient for use in a practical system.     

基于支持向量机的肝纤维化CT图像分类

基于统计学习理论中结构风险最小化原则的支持向量机是易于小样本的机器学习方法.本文使用支持向量机和二叉树的方法对肝纤维化CT图像进行分类,并与k近邻法和BP神经网络等其它算法进行比较,结果显示对于肝纤维化图像,支持向量机的分类效果和鲁棒性要高于其他两种算法.     

Correlation technique and least square support vector machine combine for frequency domain based ECG beat classification

The present work proposes the development of an automated medical diagnostic tool that can classify ECG beats. This is considered an important problem as accurate, timely detection of cardiac arrhythmia can help to provide proper medical attention to cure/reduce the ailment. The proposed scheme utilizes a cross-correlation based approach where the cross-spectral density information in frequency domain is used to extract suitable features. A least square support vector machine (LS-SVM) classifier is developed utilizing the features so that the ECG beats are classified into three categories: normal beats, PVC beats and other beats. This three-class classification scheme is developed utilizing a small training dataset and tested with an enormous testing dataset to show the generalization capability of the scheme. The scheme, when employed for 40 files in the MIT/BIH arrhythmia database, could produce high classification accuracy in the range 95.51–96.12% and could outperform several competing algorithms.     

一种基于语义模型的乳腺钙化病灶标注方法

为了实现对乳腺X线影像的医学语义标注,提出一种利用贝叶斯网络(BN)的多层乳腺影像钙化点语义建模方法。该方法首先用支持向量机(SVM)得到从图像底层视觉特征到中层特征语义的映射,然后再利用BN融合特征语义,最终提取出高层病症语义即恶性程度的概率表达,完成语义模型。将模型应用于乳腺图像的语义标注,本实验选用142幅图像作为训练集,50幅图像作为测试集,结果表明,样本标注诊断语义的准确率:恶性为81.48%,良性为73.91%。     

A comparison approach toward finding the best feature and classifier in cue-based BCI

In this paper, a comparative evaluation of state-of-the art feature extraction and classification methods is presented for five subjects in order to increase the performance of a cue-based Brain-Computer interface (BCI) system for imagery tasks (left and right hand movements). To select an informative feature with a reliable classifier features containing standard bandpower, AAR coefficients, and fractal dimension along with support vector machine (SVM), Adaboost and Fisher linear discriminant analysis (FLDA) classifiers have been assessed. In the single feature-classifier combinations, bandpower with FLDA gave the best results for three subjects, and fractal dimension and FLDA and SVM classifiers lead to the best results for two other subjects. A genetic algorithm has been used to find the best combination of the features with the aforementioned classifiers and led to dramatic reduction of the classification error and also best results in the four subjects. Genetic feature combination results have been compared with the simple feature combination to show the performance of the Genetic algorithm.     

The linker domain of poly(rC) binding protein 2 is a major determinant in poliovirus cap-independent translation

Poliovirus, a member of the enterovirus genus in the family Picornaviridae, is the causative agent of poliomyelitis. Translation of the viral genome is mediated through an internal ribosomal entry site (IRES) encoded within the 5′ noncoding region (5′ NCR). IRES elements are highly structured RNA sequences that facilitate the recruitment of ribosomes for translation. Previous studies have shown that binding of a cellular protein, poly(rC) binding protein 2 (PCBP2), to a major stem-loop structure in the genomic 5′ NCR is necessary for the translation of picornaviruses containing type I IRES elements, including poliovirus, coxsackievirus, and human rhinovirus. PCBP1, an isoform that shares approximately 90% amino acid identity to PCBP2, cannot efficiently stimulate poliovirus IRES-mediated translation, most likely due to its reduced binding affinity to stem-loop IV within the poliovirus IRES. The primary differences between PCBP1 and PCBP2 are found in the so-called linker domain between the second and third K-homology (KH) domains of these proteins. We hypothesize that the linker region of PCBP2 augments binding to poliovirus stem-loop IV RNA. To test this hypothesis, we generated six PCBP1/PCBP2 chimeric proteins. The recombinant PCBP1/PCBP2 chimeric proteins were able to interact with poliovirus stem-loop I RNA and participate in protein-protein interactions. We demonstrated that the PCBP1/PCBP2 chimeric proteins with the PCBP2 linker, but not with the PCBP1 linker, were able to interact with poliovirus stem-loop IV RNA, and could subsequently stimulate poliovirus IRES-mediated translation. In addition, using a monoclonal anti-PCBP2 antibody (directed against the PCBP2 linker domain) in mobility shift assays, we showed that the PCBP2 linker domain modulates binding to poliovirus stem-loop IV RNA via a mechanism that is not inhibited by the antibody.