A hybrid model of partial least squares and neural network for traffic incident detection

Development of a universal freeway incident detection algorithm is a task that remains unfulfilled and many promising approaches have been recently explored. The partial least squares (PLS) method and artificial neural network (NN) were found in previous studies to yield superior incident detection performance. In this article, a hybrid model which combines PLS and NN is developed to detect automatically traffic incident. A real traffic data set collected from motorways A12 in the Netherlands is presented to illustrate such an approach. Data cleansing has been introduced to preprocess traffic data sets to improve the data quality in order to increase the veracity and reliability of incident model. The detection performance is evaluated by the common criteria including detection rate, false alarm rate, mean time to detection, classification rate and the area under the curve (AUC) of the receiver operating characteristic. Computational results indicate that the hybrid approach is capable of increasing detection performance comparing to PLS, and simplifying the NN structure for incident detection. The hybrid model is a promising alternative to the usual PLS or NN for incident detection.     

Optimizing area under the ROC curve using semi-supervised learning

Receiver operating characteristic (ROC) analysis is a standard methodology to evaluate the performance of a binary classification system. The area under the ROC curve (AUC) is a performance metric that summarizes how well a classifier separates two classes. Traditional AUC optimization techniques are supervised learning methods that utilize only labeled data (i.e., the true class is known for all data) to train the classifiers. In this work, inspired by semi-supervised and transductive learning, we propose two new AUC optimization algorithms hereby referred to as semi-supervised learning receiver operating characteristic (SSLROC) algorithms, which utilize unlabeled test samples in classifier training to maximize AUC. Unlabeled samples are incorporated into the AUC optimization process, and their ranking relationships to labeled positive and negative training samples are considered as optimization constraints. The introduced test samples will cause the learned decision boundary in a multi-dimensional feature space to adapt not only to the distribution of labeled training data, but also to the distribution of unlabeled test data. We formulate the semi-supervised AUC optimization problem as a semi-definite programming problem based on the margin maximization theory. The proposed methods SSLROC1 (1-norm) and SSLROC2 (2-norm) were evaluated using 34 (determined by power analysis) randomly selected datasets from the University of California, Irvine machine learning repository. Wilcoxon signed rank tests showed that the proposed methods achieved significant improvement compared with state-of-the-art methods. The proposed methods were also applied to a CT colonography dataset for colonic polyp classification and showed promising results.¹     

求解AUC优化问题的对偶坐标下降方法

AUC被广泛作为衡量不平衡数据分类性能的评价标准.与二分类问题不同,AUC问题的损失函数由来自两个不同类别的样本对组成.如何提高其实际收敛速度,是一个值得研究的问题.目前的研究结果表明：使用reservoir sampling技术的在线方法(OAM)表现出很好的AUC性能,但OAM仍存在诸如收敛速度慢、参数选择复杂等缺点.针对 AUC 优化问题的对偶坐标下降(AUC-DCD)方法进行了系统的研究,给出3种算法,即 AUC-SDCD,AUC- SDCDperm和AUC-MSGD,其中,AUC-SDCD和AUC-SDCDperm与样本数目有关,AUC-MSGD与样本数目无关.理论分析指出,OAM是AUC-DCD的一种特殊情形.实验结果表明,AUC-DCD在AUC性能和收敛速度两方面均优于OAM.研究结果表明,AUC-DCD是求解AUC优化问题的首选方法.     

氟体系AUC流程制备陶瓷UO_2粉末及其母液处理工艺

介绍了以UF6水解液为原料采用气提环流搅拌沉淀反应器制备AUC粉末、用流化床分解还原氟体系AUC制备陶瓷级UO2粉末、用离子交换树脂处理含氟废液中的铀和用生石灰沉淀离子交换尾液中氟的处理工艺。具体研究了以UF6水解液为原料制备AUC粉末的主要控制参数;讨论了氟体系制备的AUC粉末与硝酸体系制备的粉末的各种差异。实验结果表明:采用气体环流搅拌反应器并以UF6水解液为原料制备AUC粉末,组成恒定且质量稳定,重现性好;采用流化床技术分解还原氟体系AUC晶体,并在最佳工艺条件下能够稳定地获得性能优良的陶瓷级UO2粉末,完全满足Gd2O3-UO2芯块制造的要求。     

数字调制信号识别性能的评估方法

针对正确率不能客观全面地评估字调制信号识别性能的问题,提出采用受试者操作特征(ROC)曲线下的面积(AUC)对最小二乘支持向量机分类器和传统的神经网络分类器进行性能评佑。首先提取J个特征参数,然后分别采用最小二乘支持向量机分类器和神经网络分类器成功地实现了数字调制信号识别,最后通过计算ROC曲线下的AUC值来评估分类器的优劣。仿真实验结果表明,最小二乘支持向量机分类器比神经网络分类器的平均性能好。     

Naive Bayes for optimal ranking

This paper presents experimental comparisons between the declarative encodings of various computationally hard problems in Answer Set Programming (ASP) and Constraint Logic Programming over Finite Domains (CLP(FD)). The objective is to investigate how solvers in the two domains respond to different problems, highlighting the strengths and weaknesses of their implementations, and suggesting criteria for choosing one approach over the other. Ultimately, the work in this paper is expected to lay the foundations for a transfer of technology between the two domains, for example by suggesting ways to use CLP(FD) in the execution of ASP.     

BO-AUC多类分类评估方法

分类技术是数据挖掘研究的核心技术之一,分类评估也是研究热点,基于AUC评估方法是分类评估领域的研究热点,其中B-AUC评估算法可以有效地评估分类器性能,但该评估方法有不足之处。该分类评估方法建立在不对称的两个类别上,影响了评价结果;根据非完全二叉树思想存储,浪费了存储空间;基于偏二叉树的搜索效率不高。利用完全二叉树的构造思想提出了BO-AUC评估方法,该方法将n个类别的分类问题分解为独立的二类进行成对的计算,可以有效地解决B-AUC的不足,进一步扩展基于AUC的评估标准,在MBNC实验上编程实现该方法,实验结果表明BO-AUC方法的有效性。     

Aroma impact components of rabbiteye blueberry (Vaccinium ashei) vinegars

The aroma impact components of the rabbiteye blueberry vinegars made from wine fermented without skin-contact (JV), wine fermented with skin-contact (WV), or blueberries (BV) were evaluated. Headspace volatiles were isolated by solid-phase microextraction (SPME). Headspace volatiles were analysed by gas chromatography–mass spectrometry (GC–MS) and gas chromatography–olfactometry (GCO). The potent odourants were evaluated by the ‘Osme’ method. Acetic acid, 2/3-methyl-butanoic acid, phenethyl acetate, 2-phenylethanol, octanoic acid, eugenol, and phenylacetic acid were the most important aroma-active compounds identified in all three treatments. The contributions of these compounds to each sample were different. There were more aroma-active compounds identified from the BV than from the other two treatments. The type of fermentation impacts the aroma components of rabbiteye blueberry vinegars.     

基于纹理特征的糖网临床前期眼底自发荧光图像识别

及时发现和干预潜在的糖尿病视网膜病变患者，对帮助提升糖尿病患者的整体视觉质量和降低医疗成本具有十分积极的意义.由于糖尿病视网膜病变临床前期和正常人的眼底荧光图像在视觉感观基本上没有差别，为此本文通过目前应用较广的纹理特征算法和支持向量机对这两组图像进行了模式识别.通过将185张眼底荧光图片十折交叉检验发现，LBP算法对其具有很好的识别效果.等价模式下的59维LBP算子的十折交叉准确率达到了91.89%，同时在测试集和训练集以1：1随机划分的情况下，由训练集数据所生成的模型对测试集中92张眼底荧光图像的识别准确率达到了88.12%，AUC值为0.943.