Anomaly detection based on the multi phase clustering and naive bayes

Anomaly detection method was used for calibration data concentration significantly different from other data objects. In this paper, the multi phase clustering aimed at resolving the import of noise data and the lack of the attributive sample, and changing the traditional passive learning of bayes for active learning ways to structure the superior performance classifier. In the pre processing stage, a clustering algorithm based on density is introduced to handle noise data. And the output of the density based clustering algorithm can be used as the input of K means, which responsible for handling the training samples with absent values. At classification time, we introduce adaboost algorithm into naive bayes to generate a more effective classifier.     

A Study on Urban Land Structure Optimization Model

As far as the urban land structure is mentioned, it means the distribution of urban land among the different industries of the city. The problem of urban land structure is referred as unreasonable allocation of urban land caused by several factors, including the decision making mechanism of old economy and urban development policies. Different from using Fractal theory, Entropy theory and Functional theory, in this paper, with reference to Dissipative Structure Theory, one dynamic, systematic and historical optimization model of Max f(J_i, X_i) with two parameters of J_i and X_i will mainly be set up. This model can be used for analyzing the key factors determining the optimization of urban land structure.     

Simulation of the constant modulus blind equilibrium algorithm with variable step size based on the mean square error

The mathematic models and basic theory of constant modulus algorithm (CMA) for blind channel equalization are introduced. An improved algorithm with variable step size is proposed based on CMA. The improved algorithm uses the mean square error (MSE) to obtain the new variable step size to solve the contradiction between the convergence rate and accuracy in traditional CMA with fixed step. To smooth the MSE learning curve for easy comparison, an ensemble average technique is used to analyze the performance of convergence with different step sizes. The theoretical analysis and the simulation results show that the improved algorithm is superior to the traditional constant modulus(CM) in terms of convergence speed and residual error.     

Analysis of Information Flow Based on Entropy of Fault Diagnosis

Under the pointview of epistemology and generalized information theory, fault diagnosis can be regarded as the state identification process by dint of various cognitive tools. Therefore, information Entropy can be employed to explain the transfer of information during identifying the faults. The mathematical deduction of information Entropy is given and the idea of Multi-Symtom Domains Comprehensive Feature Knowledge to solve fault diagnosis problems caused by insufficient knowledge and lower cognitiving ability for complex knowledge system is presented.     

A Self learning Expert Controller Being Fit for Controlling pH Value

In view of the characteristics of pH value,a self learning algorithm is proposed,and a self learning expert controller which being fit for controlling pH value is designed.The simulating results prove the effectiveness of the self learning algorithm.     

An Improved Learning Algorithm for BP Model and its Application to Load Forecasting in Power Systems

This paper analyses the cause of very slow learning rate during the time of ap-proaching the minim um poi nt in con vertional BP algorithm, by introducing weight factor into themodification of the connection weight,an improved leavning algorithm for BP model is presented,itcan sign ificantly in crease the learning rate and improve con vergent behaviour.As an illustrative ap-plication example,the n umeriCal resu It of the medium-term load forecasting is given.     

基于深度卷积神经网络的玉米病害识别

为了提高玉米病害的识别率,本文提出了一种在自然环境条件下基于深度卷积神经网络的玉米病害识别方法。该方法以玉米常见的10类病害为研究对象。算法模型是先将图像预处理,应用Triplet loss双卷积神经网络结构学习玉米图像特征,再使用SIFT算法提取图像纹理细节,最后通过Softmax对图像进行标签分类。训练集采用正常玉米图像与玉米病害图像相结合的方式,使用深度相似性网络学习正常玉米图像特征表示,再使用迁移学习方法学习玉米病害图像的特征,最后对特征进行分类识别。研究结果表明,该方法可准确识别10种常见玉米病害,正确率可达90%以上,为玉米病害的防治提供了有效的技术支持。     

Testing the Efficiency of Pollination Bag Materials

Crossing bags manufactured from three grades of polyester and three types of paper bags were tested for their permeability to grass pollen. Bags were placed on L. perenne plants of known PGI isozyme genotype. Any pollen passing through the crossing bag and pollinating the plant was detected by assaying its progeny for PGI isozymes. The results obtained allowed differentiation of the various materials for their suitability for pollen proofing.     

基于Memetic算法的两级车辆路径优化

针对传统两级车辆路径优化过程,存在的精度不高,收敛时间过长的问题,提出一种基于Q学习理论和差分进化的Memetic算法。首先,对两级车辆路径优化问题进行研究,利用最优分割法获得第一级配送方案,以此确定中转站配送数量,然后求解第二级多配送中心车辆路径问题配送方案,获得两级优化问题的总里程及总配送车辆数量;其次,针对第二级MDVRP配送方案求解,利用Q学习理论和差分进化算法,设计新的Memetic算法,来实现对多配送中心车辆路径问题配送方案的全局优化;最后,通过仿真验证了所提算法的有效性。     

A Virtual Wavelet Transform Analyzer for the Signal Analysis Based on the Direct Algorithm

A virtual wavelet transform analyzer for the signal analysis based on the direct algorithm is introduced so that the discrete wavelet transform and continuous wavelet transform is maken to signal in the direct algorithm. The authors first introduce the direct algorithm of the WT, which is numerical algorithm obtained from the original formula of the wavelet transform by directly numericalizing. Then some conclusions are drawn on the direct algorithm. The examples are the sampling principle and technology for the wavelets, the limitation of the scale range of the wavelets and the measures to solve the edge phenomenal in the direct algorithm of the discrete wavelet transform, and some conclusions in the direct algorithm of the continuous wavelet transform. The virtual wavelet transform analyzer for the signal analysis based on the direct algorithm explored based on these studies and combined with virtual instrument technique can make the discrete wavelet transform and continuous wavelet transform to signal with any basic wavelet. It can be applied in studying the property of any basic wavelet and learning the theory on the wavelet transform, and also in making some engineering signal analysis. In the end, the authors give some typical examples for the application of the virtual analyzer. These examples show that the analyzer can be applied in many situations.     

Algorithm to Transform Dynamic Causality Diagram Into Belief Network

The Dynamic Causality Diagram methodology is a new probabilistic reasoning model based on Belief Network. To some extent, it is similar to the Belief Network in structure. So that knowledge from one can be transformed into the other on some conditions. To begin with, this paper discusses the similarities and differences between them, and finally presents a transformation algorithm from Dynamic Causality Diagram into Belief Network. The algorithm is composed of two parts: a mapping algorithm of structure and a generating algorithm of conditional probability tables.     

陆地棉育种选择策略的分析 Ⅱ选择效果分析

本研究利用多种陆地棉种质材料,对皮棉产量及其组成性状选择效果的分析证明,铃数增加不仅因遗传力低、选择准确性差而难以获得明显进展,而且进一步改良潜力很小,极难超过对照品种。相对而言,全株平均铃重与草粒种子纤维数是两项遗传力较高,选择潜力较大,易于改进的性状。无论从性状关联性、还是从选择效果分析,目前,陆地棉高产育种都应采取稳定铃数,保持现有推广品种水平,重点提高全株平均铃重、特别是结铃与吐絮集中性和衣指的产量选择策略。     

多邻域保持嵌入的人脸识别方法

现有流形学习算法在学习人脸数据时,假设所有数据点位于单一低维嵌入流形之上,当数据点实际分布在不同的流形上时,单流形假设就会影响数据真实空间结构。为此提出一种基于多邻域保持嵌入（multiple neighborhood preserving embedding, M-NPE）的学习算法来发现不同类别数据在不同维度的低维嵌入空间中分布的多流形结构。首先,单独学习不同类别数据的流形,得到反映其本质特征的流形;再通过遗传算法搜索每个流形的最优维数;最后依据最小重构误差分类器对样本分类。在Extended Yale B和CMU PIE这2个大型人脸库上实验结果验证了该算法的有效性。     

Inheritance of spineless leaves in Aechmea (Bromeliaceae)

Bromeliaceae are tropical ornamentals with a variety of colors and inflorescence shapes. Leaves with spined margins are characteristic for the genus Aechmea. Crosses between spineless and spined Aechmea cultivars were attempted in order to obtain a spineless offspring. Crossing spined and spineless genotypes resulted in a progeny with on average 50% spineless plants. The hybrids of two spineless plants had on average 75% spineless leaves. This indicated that the spineless character was dominant. Homozygous plants were selected; these plants are interesting for further breeding and propagation by seed. This revised version was published online in July 2006 with corrections to the Cover Date.     

BP Neural Networks and Its Application in Lung Cancer Intelligent Diagnosis System

The learning algorithm of networks is discussed. The programming example of 3 layer BP networks is given with Visual C++6.0 program langue. Based on this model, a lung cancer intelligent diagnosis system is successfully implemented. Furthermore, the paper introduces network's structure design, preferencesand the source of stylebookdatum in factual applications. The ameliorative arithmetic is applied to the study of networks and BP dynamic evolving process is designed. The experiments indicatecell images are recognized and classified by the trained neural network. The study illustrates the system has feasibility and clinical value in lung cancer diagnosis.     

An improved morphological edge detection algorithm of medical image

Medical images usually contain much noise which affects the edge detection accuracy. Focusing on this problem, based on the edge detection operator in mathematical morphology, an improved edge detection algorithm is presented by combining the features of the multi structure elements and the multi scale edge detection algorithm. The algorithm performs opening and closing operations on the data with the alternative sequence filters and the structure elements. The weighting operation is applied with different weight coefficients for horizontal, vertical and diagonal directions, while the edge detection operator with dilation type is calculated to obtain the improved edge detection algorithm. The steps of the algorithm are described. The algorithm is used to extract the edge of MRI image as well as the image of Lena. The experimental results indicate the algorithm can considerably improve the edge resolution of the traditional morphological edge detection methods and is practical.     

Adaptive Iterative Learning Control of Nonlinear Systems with Unknown Control Direction

An adaptive iterative learning control scheme is proposed for a class of mismatched nonlinear systems with periodic uncertainties and unknown control directions. The control algorithm has the following three features;the control method does not need control direction information; the periodic uncertainties are learned online by a learning control method; and the algorithm can deal with mismatched uncertainties. The proposed method can achieve asymptotical convergence along a learning repetition horizon. A simulation example is provided can demonstrate the effectiveness and feasibility of the control strategy.     

Description and Breeding Behavior of Two Partial Male Sterile Guar Plants

This investigation describes two partial male sterile plants of guar, Cyamopsis tetragonoloba (L.) Taub. (MS-1 and MS-II), and gives results of their hybridization with a fertile cultivar. Detectable natural crossing between MS-1 plants and a fertile cultivar averaged 20.6%, with maximum crossing of 70 %. Crossing between MS-II plants and a fertile culuvar averaged 5.0 5.0 % about the same as previously reported for fertile × fertile, crosses. Honeybees, Apis mellifera (L.), caged with MS-II plants and a fertile cultivar did not increase seed set on male sterile plants. Hand separation of mature pods into short- and long-pod groups was only partially effective in identifying hybrid seed produced on the male sterile plants.     

A forward algorithm for prestressed construction process of large-span suspendome based on the equivalent pre-tension

To accurately simulate prestressed construction process of large-span suspendome, the initial state of construction process according to practical construction sequence is established, and the construction mechanical analysis for prestressed construction process is especially studied. Features of prestressed construction process of large-span suspendome are summarized, and the insufficiencies of the application of state variable superposition method, back analysis method, birth and death element method are analyzed. Combined with nonlinear finite element analysis method, forward algorithm for prestressed construction process based on the equivalent pre-tension is proposed, and the specific application method and procedure are given. Changzhou stadium steel roof engineering is taken as an example to verify the practicability and applicability of forward algorithm for prestressed construction process. The conclusion is that, using forward algorithm for prestressed construction process based on the equivalent pre-tension, the structure state of each construction phase can be accurately tracked, and the nonlinear contact of strut and rigid framework, interactive impact of cable force, conversion of temporary structure system, and structural geometric nonlinearity can be comprehensively considered.     

Robust Adaptive Control Based on On-line Identification to Multi-parameters

In system controlled,Sampling these parameters and with parameters for controller in time, it is an object to adaptive control. The article presents a scheme that the structure parameters on it's model and model parameters is robust identified by on line, while to optimal of algorithm on adaptive control is fulfilled. In control scheme, method of intelligent identifying to the structure parameters(levels of model) and algorithm of robust pole placement are introduced also It is advancer for already mentioned scheme. Whole scheme presents new model on adaptive control.