可交换模糊有限自动机的代数性质

研究了可交换模糊Mealy机的同态性质、可交换模糊Mealy机与其子模糊Mealy机的关系、可交换模糊Mealy机的覆盖关系。定义了可弱交换的模糊Mealy机,并将可交换的模糊Mealy机的性质推广到可弱交换的模糊Mealy机上。     

伪加权Mealy机与伪加权Moore机的关系

给出伪加权序列机、伪加权Mealy机以及伪加权Moore机的定义并分别给出了它们的响应函数.证明伪加权序列机与伪加权Mealy机是不等价的、伪加权序列机与伪加权Moore机是等价的;并以伪加权序列机为桥梁,得到了伪加权Mealy机与伪加权Moore机的关系是不等价的.     

关于DNA序列分类问题的模型

本文提出了一种将人工神经元网络用于 DNA分类的方法 .作者首先应用概率统计的方法对 2 0个已知类别的人工 DNA序列进行特征提取 ,形成 DNA序列的特征向量 ,并将之作为样本输入 BP神经网络进行学习 .作者应用了 MATLAB软件包中的 Neural Network Toolbox(神经网络工具箱 )中的反向传播 ( Backpropagation BP)算法来训练神经网络 .在本文中 ,作者构造了两个三层 BP神经网络 ,将提取的 DNA特征向量集作为样本分别输入这两个网络进行学习 .通过训练后 ,将 2 0个未分类的人工序列样本和 1 82个自然序列样本提取特征形成特征向量并输入两个网络进行分类 .结果表明 :本文中提出的分类方法能够以很高的正确率和精度对 DNA序列进行分类 ,将人工神经元网络用于 DNA序列分类是完全可行的     

矢量文法与PN机

引入串语义下的矢量文法概念,给出矢量文法的类乔姆斯基分类,并就正规矢量文法进行了更细的划分.讨论了矢量文法谱系和标量文法谱系之间的强弱关系,构成了标矢量文法谱系图。指出正规矢量文法与Petri网(也称PN机)语言上的等同关系,引入混杂PN机,并证明其语言与上下文无关矢量文法的语言是等同的.由此部分构成了矢量文法与PN机之间的关系结构.     

基于Kohonen网络的激光超声缺陷探测研究

在激光超声缺陷检测技术中,不同类型缺陷采样信号的准确分类至关重要.针对激光超声表面波实验采样信号高维小样本的特点,采用了一种有监督学习的Kohonen神经网络(S_Kohonen)自适应分类方法.在S_Kohonen网络自组织学习的过程中,通过改进网络的学习率提高了网络的收敛速度.通过采用一种无需邻域半径判断的自适应权值调整方式来实现竞争层神经元权值不同程度的调整,从而更有效的表征输入样本的分布特征.通过对不同类型缺陷探测样本的多次实验,验证了所述方法具有良好的分类预测效果,多次交叉验证分类正确率均能达到100%.     

高斯核正则化学习算法的泛化误差

对广义凸损失函数和变高斯核情形下正则化学习算法的泛化性能展开研究.其目标是给出学习算法泛化误差的一个较为满意上界.泛化误差可以利用正则误差和样本误差来测定.基于高斯核的特性,通过构构建一个径向基函数(简记为RBF)神经网络,给出了正则误差的上界估计,通过投影算子和再生高斯核希尔伯特空间的覆盖数给出样本误差的上界估计.所获结果表明,通过适当选取参数σ和λ,可以提高学习算法的泛化性能.     

智能制造模式下基于改进BP-ARIMA组合模型产品需求预测方法

为有效预测智能制造模式下的不确定性需求,提出自回归移动平均模型ARIMA和改进BP神经网络的组合模型,对预测数据中包含线性规律的Lt以及非线性规律的ε_t进行模拟和分析,以解决预测有效性和精度问题.通过数据样本构建,对ARIMA模型结构进行辨识,确定p,d,q参数,并对模型进行诊断和检验;在此基础上进行需求数据一次预测;通过连接权值的修正降低BP神经网络学习误差,并对一次预测结果与原需求数据样本存在的误差进行二次预测.实例数据分析表明:组合模型的预测精度较ARIMA模型有显著提高,因此组合预测模型在预测效果上具有合理性和有效性.     

数据挖掘在房地产价格预测中的应用研究

通过基于数据挖掘理论的粗糙集和神经网络的研究,用属性约简算法约简并提取了影响房地产价格的主要指标因素,对降维后的数据进行网络学习和训练,最后用训练好的的网络检验测试样本.方法使学习训练的速度和识别率提高了,为房地产价格预测提供了一种更为有效和实用的新途径.     

零样本学习进展研究

机器学习是人工智能领域发展最迅速的一个分支之一,传统的机器学习方法和深度学习大都需要大量人工标注的训练数据才能发挥作用.然而,现实世界的物体种类繁多且其数量在不断增长,人工标注训练数据就变成了一项极其繁琐冗杂的工作,零样本学习的提出极大地缓解了这种情况.在零样本学习中,训练集和测试集的类别的交集是空集,因此需要在二者之间通过实现知识的迁移来完成学习,从而使得在训练集上训练得到的模型能够识别测试集上输入示例的类别标签.不同于其他大部分机器学习技术需要保证训练集包含测试集,零样本学习的原理从本质意义上让计算机模仿了人类在学习时的推理模式,使得计算机能够识别新事物.本文梳理了零样本学习的研究进展,首先概述了零样本学习的定义及其相关领域,然后重点归纳了零样本学习的发展过程,包括其基本模型及改进,存在的关键难点以及解决方式,最后探讨了零样本学习的研究现状及其未来的发展方向.     

基于模糊字符串的Mealy格值有限自动机及其最小化

提出取值为格半群的Mealy格值有限自动机的概念,进而得到基于模糊字符串的Mealy格值有限自动机的扩张模型,并较详细讨论了其性质. 同时定义了扩张的完备Mealy格值有限自动机的行为矩阵, 在此基础上给出了其最小化算法.     

On the synthesis of finite sequential machines

Starting from the studies of Kleene and Mealy on sequential machines, in this paper is presented a formalism which, in a sense, unifies their treatments. From the specification of the required machine behaviour in terms of events and associated output states, a uniform procedure is given for obtaining a transition table and from that a minimal machine, whenever such a complete reduction is possible. The various steps of the synthesis procedure are so stated that they can be easily programmed on a computer.     

分类问题的在线顺序双并联快速学习机

双并联前馈神经网络模型是单层感知机和单隐层前馈神经网络的混合结构,本文构造了一种双并联快速学习机算法,与其他类似算法比较,提出的算法能利用较少的隐层单元及更少的待定参数,获得近似的学习性能.数值实验表明,对很多实际分类问题,提出的算法具备更佳的泛化能力,因而可以作为快速学习机算法的有益补充.     

Universality of deep convolutional neural networks

Deep learning has been widely applied and brought breakthroughs in speech recognition, computer vision, and many other domains. Deep neural network architectures and computational issues have been well studied in machine learning. But there lacks a theoretical foundation for understanding the approximation or generalization ability of deep learning methods generated by the network architectures such as deep convolutional neural networks. Here we show that a deep convolutional neural network (CNN) is universal, meaning that it can be used to approximate any continuous function to an arbitrary accuracy when the depth of the neural network is large enough. This answers an open question in learning theory. Our quantitative estimate, given tightly in terms of the number of free parameters to be computed, verifies the efficiency of deep CNNs in dealing with large dimensional data. Our study also demonstrates the role of convolutions in deep CNNs.     

Machine-part cell formation in group technology using a modified ART1 method

Group Technology (GT) is a useful way of increasing the productivity for manufacturing high quality products and improving the flexibility of manufacturing systems. Cell formation (CF) is a key step in GT. It is used in designing good cellular manufacturing systems using the similarities between parts in relation to the machines in their manufacture. It can identify part families and machine groups. Recently, neural networks (NNs) have been widely applied in GT due to their robust and adaptive nature. NNs are very suitable in CF with a wide variety of real applications. Although Dagli and Huggahalli adopted the ART1 network with an application in machine-part CF, there are still several drawbacks to this approach. To address these concerns, we propose a modified ART1 neural learning algorithm. In our modified ART1, the vigilance parameter can be simply estimated by the data so that it is more efficient and reliable than Dagli and Huggahalli’s method for selecting a vigilance value. We then apply the proposed algorithm to machine-part CF in GT. Several examples are presented to illustrate its efficiency. In comparison with Dagli and Huggahalli’s method based on the performance measure proposed by Chandrasekaran and Rajagopalan, our modified ART1 neural learning algorithm provides better results. Overall, the proposed algorithm is vigilance parameter-free and very efficient to use in CF with a wide variety of machine/part matrices.     

Topology control in the mobile ad hoc networks in order to intensify energy conservation

Although studied for years, due to their dynamic nature, research in the field of mobile ad hoc networks (MANETs) has remained a vast area of interest. Since once distributed, there will be less to no plausibility of recharge, energy conservation has become one of the pressing concerns regarding this particular type of network. In fact, one of the main obligations of designers is to make efficient use of these scarce resources. There has been tremendous work done in different layers of protocol stack in order to intensify energy conservation. To date, numerous topology control algorithms have been proposed, however, only a few have used meta-heuristics such as genetic algorithms, neural networks and/or learning automata to overcome this issue. On the other hand, since nodes are mobile and thus in a different spatial position, as time varies, we can expect that by regulating time intervals between topology controls, one may prolong the network’s lifetime. The main initiative of this paper is to intensify energy conservation in a mobile ad hoc network by using weighted and learning automata based algorithms. The learning automata, regulates time intervals between which the topology controls are done. The represented learning automata based algorithm uses its learning ability to find appropriate time-intervals so that the nodes would regulate the energy needed in order to exchange the information to their neighbors, accordingly. Moreover, at first we have represented two weighted based algorithms which extend two prominent protocols, namely K-Neigh and LMST. Then these algorithms are combined with a learning based algorithm which regulates time intervals between which the topology controls are done. In comparison with approaches that are based on periodic topology controls, proposed approach shows enhanced results. On the other hand, considering the learning ability of the learning automata based algorithms, composition of the aforementioned algorithms has been proven to be enhanced, in the respect of energy consumed per data transmitted, over those compared with.     

模糊无限状态自动机及其收敛性

到目前为止,我们所研究的模糊或非模糊的自动机都是有限状态自动机.然而,关于无限状态自动机的定义及它的稳定性和收敛性都没有被讨论过.本文中,我们使用离散的反馈神经网络及网络输出空间划分方法,同时,在梯度更新算法中使用伪梯度方法,给出了模糊无限状态自动机收敛到模糊有限状态自动机的证明.     

Architecture for development of adaptive on-line prediction models

This work presents an architecture for the development of on-line prediction models. The architecture defines unified modular environment based on three concepts from machine learning, these are: (i) ensemble methods, (ii) local learning, and (iii) meta learning. The three concepts are organised in a three layer hierarchy within the architecture. For the actual prediction making any data-driven predictive method such as artificial neural network, support vector machines, etc. can be implemented and plugged in. In addition to the predictive methods, data pre-processing methods can also be implemented as plug-ins. Models developed according to the architecture can be trained and operated in different modes. With regard to the training, the architecture supports the building of initial models based on a batch of training data, but if this data is not available the models can also be trained in incremental mode. In a scenario where correct target values are (occasionally) available during the run-time, the architecture supports life-long learning by providing several adaptation mechanisms across the three hierarchical levels. In order to demonstrate its practicality, we show how the issues of current soft sensor development and maintenance can be effectively dealt with by using the architecture as a construction plan for the development of adaptive soft sensing algorithms.     

偏最小二乘回归的贝叶斯正则化神经网络集成模型在证券分析预测中的应用

神经网络集成技术能有效地提高神经网络的预测精度和泛化能力,已经成为机器学习和神经计算领域的一个研究热点.利用Bagging技术和不同的神经网络算法生成集成个体,并用偏最小二乘回归方法从中提取集成因子,再利用贝叶斯正则化神经网络对其集成,以此建立上证指数预测模型.通过上证指数开、收盘价进行实例分析,计算结果表明该方法预测精度高、稳定性好.     

加权支持向量机在证券指数预测中的研究

支持向量机是基于统计学习理论的新一代学习机器.它使用结构风险最小化原则,运用核技巧,较好地解决了学习问题.本文提出了一种基于支持向量机的加权算法,并将其应用于证券,指数预测.与径向基神经网络相比较,加权支持向量机表现出了良好的性能.