Genetically optimized fuzzy polynomial neural networks with fuzzy set-based polynomial neurons

In this paper, we propose and investigate a new category of neurofuzzy networks—fuzzy polynomial neural networks (FPNN) endowed with fuzzy set-based polynomial neurons (FSPNs) We develop a comprehensive design methodology involving mechanisms of genetic optimization, and genetic algorithms (GAs) in particular. The conventional FPNNs developed so far are based on the mechanisms of self-organization, fuzzy neurocomputing, and evolutionary optimization. The design of the network exploits the FSPNs as well as the extended group method of data handling (GMDH). Let us stress that in the previous development strategies some essential parameters of the networks (such as the number of input variables, the order of the polynomial, the number of membership functions, and a collection of the specific subset of input variables) being available within the network are provided by the designer in advance and kept fixed throughout the overall development process. This restriction may hamper a possibility of developing an optimal architecture of the model. The design proposed in this study addresses this issue. The augmented and genetically developed FPNN (gFPNN) results in a structurally optimized structure and comes with a higher level of flexibility in comparison to the one we encounter in the conventional FPNNs. The GA-based design procedure being applied at each layer of the FPNN leads to the selection of the most suitable nodes (or FSPNs) available within the FPNN. In the sequel, two general optimization mechanisms are explored. First, the structural optimization is realized via GAs whereas the ensuing detailed parametric optimization is carried out in the setting of a standard least square method-based learning. The performance of the gFPNN is quantified through experimentation in which we use a number of modeling benchmarks—synthetic and experimental data being commonly used in fuzzy or neurofuzzy modeling. The obtained results demonstrate the superiority of the proposed networks over the models existing in the references.     

A critical review of the most popular types of neuro control

In this review article, the most popular types of neural network control systems are briefly introduced and their main features are reviewed. Neuro control systems are defined as control systems in which at least one artificial neural network (ANN) is directly involved in generating the control command. Initially, neural networks were mostly used to model system dynamics inversely to produce a control command which pushes the system towards a desired or reference value of the output (1989). At the next stage, neural networks were trained to track a reference model, and ANN model reference control appeared (1990). In that method, ANNs were used to extend the application of adaptive reference model control, which was a well‐known control technique. This attitude towards the extension of the application of well‐known control methods using ANNs was followed by the development of ANN model‐predictive (1991), ANN sliding mode (1994) and ANN feedback linearization (1995) techniques. As the first category of neuro controllers, inverse dynamics ANN controllers were frequently used to form a control system together with other controllers, but this attitude faded as other types of ANN control systems were developed. However, recently, this approach has been revived. In the last decade, control system designers started to use ANNs to compensate/cancel undesired or uncertain parts of systems' dynamics to facilitate the use of well‐known conventional control systems. The resultant control system usually includes two or three controllers. In this paper, applications of different ANN control systems are also addressed. Copyright © 2011 John Wiley and Sons Asia Pte Ltd and Chinese Automatic Control Society     

数据挖掘预测模型在脑损伤患者认知功能康复中的应用

为了更好地预测后天性脑损伤ABI(Acquired Brain Injury)患者认知功能康复的影响因素,提出基于决策树(DT)、多层感知器(MLP)和广义回归神经网络(GRNN)的三种预测模型。借助于10折交叉验证测试算法,通过专一性、灵敏度和精度分析以及混淆矩阵分析对模型的性能进行测试,从而获得新的知识以评估和改善认知功能康复过程中的有效性。实验结果表明,基于DT的模型的模拟结果明显比其他模型更为优越,预测平均精度可高达90.38%。     

利用模糊感知器解决模糊可分样本的分类问题

针对一类用于解决分类问题的模糊感知器,提出完全随机输入的模糊δ-规则,并给出训练样本模糊可分的定义。实例表明,利用该算法可以有效地解决模糊可分样本的分类问题,在有限步迭代后就达到收敛,即有限步训练后网络能将所有样本正确分类。     

融合评分矩阵和评论文本深度神经网络推荐模型

针对个性化推荐中用户评分矩阵数据集稀疏,用户和项目描述信息未充分利用的问题,提出融合评分矩阵和评论文本的深度神经网络推荐模型(deep neural network recommendation model,DeepRec)。首先将通过数据预处理得到的用户偏好特征和项目属性特征的文本集合分别输入到卷积神经网络进行训练,得到用户和项目的深层次非线性特征,同时将评分矩阵输入多层感知机得到用户偏好隐表示,并对两种模型提取的用户偏好隐表示进行融合;其次利用多层感知机建模用户和项目隐表示对用户进行个性化推荐;最后基于三组数据集以均方根误差为评估指标进行对比实验。结果表明DeepRec的预测误差更低,有效提高了推荐的精准度。     

基于MLP和SPSO的机器人行为选择与运动控制方法

针对传统行为选择机制(ASM)不能很好地做出控制决策的问题,提出一种基于多层感知(MLP)前馈神经网络的ASM,并将其应用到移动机器人目标跟踪中。首先,根据具体应用场景预定义多个机器人行为。然后,根据机器人配备的图像和红外传感器获得的目标位置和障碍物信息,通过MLP神经网络从预定义行为中选择出所需执行的行为。另外,为了构造最优的MLP模型,采用一种简化粒子群算法(SPSO)来优化网络权值参数。机器人目标跟踪仿真的结果表明,提出的ASM能够准确选择出合适的行为,实现了控制机器人跟踪目标移动且能够避开各种障碍物。     

Location of steel reinforcement in concrete using ground penetrating radar and neural networks

Ground-penetrating radar is becoming increasingly popular for use as a non-destructive assessment method for investigating reinforced concrete structures. The amount of data collected however can be very large and take a significant level of subjective experience to interpret. This study focuses upon the use of a neural network approach to automate and facilitate the post-processing of ground penetrating radar results. The radar data is reduced to a simplified data set by using an edge detection routine. Signal reflections from reinforcing bars displaying a hyperbolic image format are detected using a multi-layer perceptron (MLP) network with a single hidden layer containing 8 nodes to recognise a simplified hyperbolic shape. Training and testing of the network was carried out making use of an emulsion analogue tank, simulating the properties of concrete, and using real concrete specimens. The results showed that the use of a MLP neural network approach could be quite effective in automating the identification and location of embedded steel reinforcing bars from a radar investigation. Accurate estimation of depth, or cover, requires a reliable knowledge of the dielectric properties of the concrete, and recent work using a specially-developed wideband horn antenna for direct determination of in situ properties is also outlined.     

Neural networks in process life cycle profit modelling

Changes in operational environment of the process industry such as decreasing selling prices, increased competition between companies and new legislation, set requirements for performance and effectiveness of the industrial production lines and processes. For the basis of this study, a life cycle profit (LCP) model of a pulp process was constructed using different kind of process information including chemical consumptions and production levels of material and energy flows in unit processes. However, all the information needed in the creation of relevant LCP model was not directly provided by information systems of the plant. In this study, neural networks was used to model pulp bleaching process and fill out missing information and furthermore to create estimators for the alkaline chemical consumption. A data-based modelling approach was applied using an example, where factors affecting the sodium hydroxide consumption in the bleaching stage were solved. The results showed that raw process data can be refined into new valuable information using computational methods and moreover to improve the accuracy of life cycle profit models.     

面向 PDR 定位系统的多层感知机 IMU 阵列数据融合方法

为提高低成本惯性测量单元(intertial measurement unit,IMU)阵列的行人航位推算(pedestrian dead reckoning,PDR)定位 精度,首次提出了采用多层感知机(multi-layer perceptron,MLP)实现低成本 IMU 阵列数据融合的算法,通过将自主设计的 IMU 阵列和高精度 IMU 同步运动来获得 IMU 阵列的测量数据(包括三轴加速度和三轴角速度)和高精度 IMU 的测量数据,以高精 度 IMU 的测量数据作为标签,利用 MLP 将 IMU 阵列的测量数据融合,预测出物体的实际加速度和角速度,并用定位算法进行 验证。 在定位实验中,使用 MLP 融合后的预测数据的 PDR 定位精度比使用单个 IMU 测量数据的 PDR 定位精度提高了 33. 9%;比使用简单平均处理的 IMU 阵列测量数据的 PDR 定位精度提高了 20. 8%;比使用最小二乘法融合的 IMU 阵列测量数 据的 PDR 定位精度提高了 11. 6%,证明了本文所提出方法的可行性和有效性。     

基于MLP神经网络的圆锥角膜辅助诊断北大核心CSCD

圆锥角膜在病变过程中会导致角膜中央部位向前凸出,使角膜呈现出圆锥形,而且会导致高度不规则近视和散光,对视力造成不同程度损害。疾病一般发生于青少年时期,为了能及时治疗避免病变严重,筛查区分圆锥角膜具有十分重要的意义。而且临床上对于圆锥角膜诊断通常是采用角膜地形图的方法,可以得到角膜形态学的改变,但是有一定的误诊率。目前研究发现,角膜力学特性改变先于形态学,所以本文从角膜生物力学角度出发,提出一种基于多层感知机(multi-layer perceptron,MLP)神经网络区分圆锥角膜的模型。首先,利用可视化生物力学分析仪(corneal visualization scheimpflug technology,Corvis-ST)测得角膜的生物力学视频,进行处理计算得到角膜生物力学参数作为数据集,其中包含正常角膜和圆锥角膜2种类别;然后,针对角膜生物力学参数数据集构建MLP神经网络模型,将70%数据集作为训练集,30%数据集作为测试集。在数据集上训练及测试的结果表明,该模型区分圆锥角膜的准确率为97.6%。