分类错误指导的分层B-CNN模型用于细粒度分类

目的 细粒度分类近年来受到了越来越多研究者的广泛关注,其难点是分类目标间的差异非常小。为此提出一种分类错误指导的分层双线性卷积神经网络模型。方法 该模型的核心思想是将双线性卷积神经网络算法（B-CNN）容易分错、混淆的类再分别进行重新训练和分类。首先,为得到易错类,提出分类错误指导的聚类算法。该算法基于受限拉普拉斯秩（CLR）聚类模型,其核心“关联矩阵”由“分类错误矩阵”构造。其次,以聚类结果为基础,构建了新的分层B-CNN模型。结果 用分类错误指导的分层B-CNN模型在CUB-200-2011、 FGVC-Aircraft-2013b和Stanford-cars 3个标准数据集上进行了实验,相比于单层的B-CNN模型,分类准确率分别由84.35%,83.56%,89.45%提高到了84.67%,84.11%,89.78%,验证了本文算法的有效性。结论 本文提出了用分类错误矩阵指导聚类从而进行重分类的方法,相对于基于特征相似度而构造的关联矩阵,分类错误矩阵直接针对分类问题,可以有效提高易混淆类的分类准确率。本文方法针对比较相近的目标,尤其是有非常相近的目标的情况,通过将容易分错、混淆的目标分组并进行再训练和重分类,使得分类效果更好,适用于细粒度分类问题。     

Support Vector Machines for TCP traffic classification

Support Vector Machines (SVM) represent one of the most promising Machine Learning (ML) tools that can be applied to the problem of traffic classification in IP networks. In the case of SVMs, there are still open questions that need to be addressed before they can be generally applied to traffic classifiers. Having being designed essentially as techniques for binary classification, their generalization to multi-class problems is still under research. Furthermore, their performance is highly susceptible to the correct optimization of their working parameters. In this paper we describe an approach to traffic classification based on SVM. We apply one of the approaches to solving multi-class problems with SVMs to the task of statistical traffic classification, and describe a simple optimization algorithm that allows the classifier to perform correctly with as little training as a few hundred samples. The accuracy of the proposed classifier is then evaluated over three sets of traffic traces, coming from different topological points in the Internet. Although the results are relatively preliminary, they confirm that SVM-based classifiers can be very effective at discriminating traffic generated by different applications, even with reduced training set sizes.     

Affect classification using genetic-optimized ensembles of fuzzy ARTMAPs

Training neural networks in distinguishing different emotions from physiological signals frequently involves fuzzy definitions of each affective state. In addition, manual design of classification tasks often uses sub-optimum classifier parameter settings, leading to average classification performance. In this study, an attempt to create a framework for multi-layered optimization of an ensemble of classifiers to maximize the system's ability to learn and classify affect, and to minimize human involvement in setting optimum parameters for the classification system is proposed. Using fuzzy adaptive resonance theory mapping (ARTMAP) as the classifier template, genetic algorithms (GAs) were employed to perform exhaustive search for the best combination of parameter settings for individual classifier performance. Speciation was implemented using subset selection of classification data attributes, as well as using an island model genetic algorithms method. Subsequently, the generated population of optimum classifier configurations was used as candidates to form an ensemble of classifiers. Another set of GAs were used to search for the combination of classifiers that would result in the best classification ensemble accuracy. The proposed methodology was tested using two affective data sets and was able to produce relatively small ensembles of fuzzy ARTMAPs with excellent affect recognition accuracy.     

Operator functional state classification using multiple psychophysiological features in an air traffic control task

We studied 2 classifiers to determine their ability to discriminate among 4 levels of mental workload during a simulated air traffic control task using psychophysiological measures. Data from 7 air traffic controllers were used to train and test artificial neural network and stepwise discriminant classifiers. Very high levels of classification accuracy were achieved by both classifiers. When the 2 task difficulty manipulations were tested separately, the percentage correct classifications were between 84% and 88%. Feature reduction using saliency analysis for the artificial neural networks resulted in a mean of 90% correct classification accuracy. Considering the data as a 2-class problem, acceptable load versus overload, resulted in almost perfect classification accuracies, with mean percentage correct of 98%. In applied situations, the most important distinction among operator functional states would be to detect mental overload situations. These results suggest that psychophysiological data are capable of such discriminations with high levels of accuracy. Potential applications of this research include test and evaluation of new and modified systems and adaptive aiding.     

Rules extraction from constructively trained neural networks based on genetic algorithms

The application of neural networks in the data mining has become wider. Although neural networks may have complex structure, long training time, and the representation of results is not comprehensible, neural networks have high acceptance ability for noisy data, high accuracy and are preferable in data mining. On the other hand, It is an open question as to what is the best way to train and extract symbolic rules from trained neural networks in domains like classification. In this paper, we train the neural networks by constructive learning and present the analysis of the convergence rate of the error in a neural network with and without threshold which have been learnt by a constructive method to obtain the simple structure of the network.The response of ANN is acquired but its result is not in understandable form or in a black box form. It is frequently desirable to use the model backwards and identify sets of input variable which results in a desired output value. The large numbers of variables and nonlinear nature of many materials models that can help finding an optimal set of difficult input variables. We will use a genetic algorithm to solve this problem. The method is evaluated on different public-domain data sets with the aim of testing the predictive ability of the method and compared with standard classifiers, results showed comparatively high accuracy.     

Comparison of two different PNN training approaches for satellitecloud data classification

Presents a training algorithm for probabilistic neural networks (PNN) using the minimum classification error (MCE) criterion. A comparison is made between the MCE training scheme and the widely used maximum likelihood (ML) learning on a cloud classification problem using satellite imagery data.     

基于MCE训练算法的说话人辨认系统

提出了一种基于最小分类错误准则的概率神经网络的训练算法。实验结果表明，该系统及其MCE学习算法在20个说话人辨认应用中利用5s清晰语音获得98．9％的辨认率，利用l5s电话语音获得86．2％的辨认率。     

Entity identification for heterogeneous database integration—a multiple classifier system approach and empirical evaluation

Entity identification, i.e., detecting semantically corresponding records from heterogeneous data sources, is a critical step in integrating the data sources. The objective of this research is to develop and evaluate a novel multiple classifier system approach that improves entity identification accuracy. We apply various classification techniques drawn from statistical pattern recognition, machine learning, and artificial neural networks to determine whether two records from different data sources represent the same real-world entity. We further employ a variety of ways to combine multiple classifiers for improved classification accuracy. In this paper, we report on some promising empirical results that demonstrate performance improvement by combining multiple classifiers.     

A new ensemble feature selection and its application to pattern classification

Neural network ensemble based on rough sets reduct is proposed to decrease the computational complexity of conventional ensemble feature selection algorithm. First, a dynamic reduction technology combining genetic algorithm with resampling method is adopted to obtain reducts with good generalization ability. Second, Multiple BP neural networks based on different reducts are built as base classifiers. According to the idea of selective ensemble, the neural network ensemble with best generalization ability can be found by search strategies. Finally, classification based on neural network ensemble is implemented by combining the predictions of component networks with voting. The method has been verified in the experiment of remote sensing image and five UCI datasets classification. Compared with conventional ensemble feature selection algorithms, it costs less time and lower computing complexity, and the classification accuracy is satisfactory.     

A machine learning approach for feature selection traffic classification using security analysis

Class imbalance has become a big problem that leads to inaccurate traffic classification. Accurate traffic classification of traffic flows helps us in security monitoring, IP management, intrusion detection, etc. To address the traffic classification problem, in literature, machine learning (ML) approaches are widely used. Therefore, in this paper, we also proposed an ML-based hybrid feature selection algorithm named WMI_AUC that make use of two metrics: weighted mutual information (WMI) metric and area under ROC curve (AUC). These metrics select effective features from a traffic flow. However, in order to select robust features from the selected features, we proposed robust features selection algorithm. The proposed approach increases the accuracy of ML classifiers and helps in detecting malicious traffic. We evaluate our work using 11 well-known ML classifiers on the different network environment traces datasets. Experimental results showed that our algorithms achieve more than 95% flow accuracy results.     

基于双池化特征加权结构CNN的图像分类

传统的池化方式会造成特征信息丢失,导致卷积神经网络中提取的特征信息不足。为了提高卷积神经网络在图像分类过程中的准确率,优化其学习性能,本文在传统池化方式的基础上提出一种双池化特征加权结构的池化算法,利用最大池化和平均池化2种方式保留更多的有价值的特征信息,并通过遗传算法对模型进行优化。通过训练不同池化方式的卷积神经网络,研究卷积神经网络在不同数据集上的分类准确率和收敛速度。实验在遥感图像数据集NWPU-RESISC45和彩色图像数据集Cifar-10上对采用几种池化方式的卷积神经网络分类结果进行对比验证,结果分析表明：双池化特征加权结构使得卷积神经网络的分类准确率有很大程度的提高,同时模型的收敛速度得到进一步提高。     

非平衡网络流量识别方法

针对网络中存在的对等网络（P2P）流量泛滥导致的流量失衡问题,提出将非平衡数据分类思想应用于流量识别过程。通过引入合成少数类过采样技术（SMOTE）算法并进行改进,提出了均值SMOTE （M-SMOTE）算法,实现对流量数据的平衡化处理。在此基础上分别采用3种机器学习分类器：随机森林（RF）、支持向量机（SVM）、反向传播神经网络（BPNN）对处理后各类流量进行识别。理论分析与仿真结果表明,在不影响P2P流量识别准确率的前提下,与非平衡状态相比,引入SMOTE算法将非P2P流量的识别准确率平均提高了16.5个百分点,将网络流量的整体识别率提高了9.5个百分点;与SMOTE算法相比,M-SMOTE算法将非P2P流量的识别准确率与网络流量的整体识别率分别进一步提高了3.2个百分点和2.6个百分点。实验结果表明,非平衡数据分类思想可有效解决P2P流量过多导致的非P2P流量识别率低的问题,同时所提M-SMOTE算法具有更高的识别准确度。     

一种基于优化的RBF神经网络交通流预测新算法

提出了一种基于遗传算法优化的RBF神经网络交通流预测新方法,该方法把遗传算法应用于RBF神经网络的参数确定中,实现了RBF神经网络隐层高斯函数的中心矢量和基宽向量以及隐层与输出层之间的权值的优化,提高了RBF神经网络的泛化能力。仿真结果表明：改进的RBF网络用于交通流预测中具有可靠的精度和较好的收敛速度,具有广阔的应用推广前景。     

An investigation into the application of neural networks,fuzzy logic,genetic algorithms,and rough sets to automated knowledge acquisition for classification problems

This paper presents some highlights in the application of neural networks, fuzzy logic, genetic algorithms, and rough sets to automated knowledge acquisition. These techniques are capable of dealing with inexact and imprecise problem domains and have been demonstrated to be useful in the solution of classification problems. It addresses the issue of the application of appropriate evaluation criteria such as rule base accuracy and comprehensibility for new knowledge acquisition techniques. An empirical study is then described in which three approaches to knowledge acquisition are investigated. The first approach combines neural networks and fuzzy logic, the second, genetic algorithms and fuzzy logic, and in the third a rough sets approach has been examined, and compared. In this study neural network and genetic algorithm fuzzy rule induction systems have been developed and applied to three classification problems. Rule induction software based on rough sets theory was also used to generate and test rule bases for the same data. A comparison of these approaches with the C4.5 inductive algorithm was also carried out. Our research to date indicates that, based on the evaluation criteria used, the genetic/fuzzy approach compares more than favourably with the neuro/fuzzy and rough set approaches. On the data sets used the genetic algorithm system displays a higher accuracy of classification and rule base comprehensibility than the C4.5 inductive algorithm.     

基于改进的旋转森林算法的不平衡网络流量分类方法

针对不平衡网络流量分类精度不高的问题,在旋转森林算法的基础上结合Bagging算法的Bootstrap抽样和基于分类精度排序的基分类器选择算法,提出一种改进的旋转森林算法。首先,对原始训练集按特征进行子集划分并分别使用Bagging进行样本抽样,通过主成分分析(PCA)生成主成分系数矩阵;然后,在原始训练集和主成分系数矩阵的基础上进行特征转换,生成新的训练子集,再次使用Bagging对子集进行抽样,提升训练集的差异性,并使用训练子集训练C4.5基分类器;最后,使用测试集评价基分类器,依据总体分类精度进行排序筛选,保留分类精度较高的分类器并生成一致分类结果。在不平衡网络流量数据集上进行测试实验,依据准确率和召回率两个标准对C4.5、Bagging、旋转森林和改进的旋转森林四种算法评价,依据模型训练时间和测试时间评价四种算法的时间效率。实验结果表明改进的旋转森林算法对万维网(WWW)协议、Mail协议、Attack协议、对等网(P2P)协议的分类准确度达到99.5%以上,召回率也高于旋转森林、Bagging、C4.5三种算法,可用于网络入侵取证、维护网络安全、提升网络服务质量。     

Selecting salient features for classification based on neural network committees

Aggregating outputs of multiple classifiers into a committee decision is one of the most important techniques for improving classification accuracy. The issue of selecting an optimal subset of relevant features plays also an important role in successful design of a pattern recognition system. In this paper, we present a neural network based approach for identifying salient features for classification in neural network committees. Feature selection is based on two criteria, namely the reaction of the cross-validation data set classification error due to the removal of the individual features and the diversity of neural networks comprising the committee. The algorithm developed removed a large number of features from the original data sets without reducing the classification accuracy of the committees. The accuracy of the committees utilizing the reduced feature sets was higher than those exploiting all the original features.     

Integration of independent component analysis and neural networks for ECG beat classification

In this paper, we propose a scheme to integrate independent component analysis (ICA) and neural networks for electrocardiogram (ECG) beat classification. The ICA is used to decompose ECG signals into weighted sum of basic components that are statistically mutual independent. The projections on these components, together with the RR interval, then constitute a feature vector for the following classifier. Two neural networks, including a probabilistic neural network (PNN) and a back-propagation neural network (BPNN), are employed as classifiers. ECG samples attributing to eight different beat types were sampled from the MIT-BIH arrhythmia database for experiments. The results show high classification accuracy of over 98% with either of the two classifiers. Between them, the PNN shows a slightly better performance than BPNN in terms of accuracy and robustness to the number of ICA-bases. The impressive results prove that the integration of independent component analysis and neural networks, especially PNN, is a promising scheme for the computer-aided diagnosis of heart diseases based on ECG.     

基于多进化神经网络的信用评估模型研究

客户信用评佑对于银行的经营管理有着重要的意义,为此提出了一种基于多进化神经网络的信用评估模型(MNN-CREDIT)。该模型基于客户信货数据,利用基于聚类的小生境遗传算法并行地训练出多个精度高、差异性大的三层前馈神经网络,然后将待识别的客户数据分别输入,最后根据动态投票法集成最终信用预测结果。利用德国信用数据库真实数据集进行了实证分析,结果表明,基于多进化神经网络的信用评估模型具有较高的预测精度。     

一种基于多进化神经网络的分类方法

分类问题是目前数据挖掘和机器学习领域的重要内容.提出了一种基于多进化神经网络的分类方法CABEN(classification approach based on evolutionary neural networks).利用改进的进化策略和Levenberg-Marquardt方法对多个三层前馈神经网络同时进行训练.训练好各个分类模型以后,将待识别数据分别输入,最后根据绝对多数投票法决定最终分类结果.实验结果表明,该方法可以较好地进行数据分类,而且与传统的神经网络方法以及贝叶斯方法和决策树方法相比,在     

Probabilistic neural network for breast cancer classification

Among cancers, breast cancer causes second most number of deaths in women. To reduce the high number of unnecessary breast biopsies, several computer-aided diagnosis systems have been proposed in the last years. These systems help physicians in their decision to perform a breast biopsy on a suspicious lesion seen in a mammogram or to perform a short-term follow-up examination instead. In clinical diagnosis, the use of artificial intelligent techniques as neural networks has shown great potential in this field. In this paper, three classification algorithms, multi-layer perceptron (MLP), radial basis function (RBF) and probabilistic neural networks (PNN), are applied for the purpose of detection and classification of breast cancer. Decision making is performed in two stages: training the classifiers with features from Wisconsin Breast Cancer database and then testing. The performance of the proposed structure is evaluated in terms of sensitivity, specificity, accuracy and ROC. The results revealed that PNN was the best classifiers by achieving accuracy rates of 100 and 97.66 % in both training and testing phases, respectively. MLP was ranked as the second classifier and was capable of achieving 97.80 and 96.34 % classification accuracy for training and validation phases, respectively, using scaled conjugate gradient learning algorithm. However, RBF performed better than MLP in the training phase, and it has achieved the lowest accuracy in the validation phase.