Description: 一种基于蚁群聚类的径向基神经网络
提出了一种基于蚁群聚类算法的径向基神经网络. 利用蚁群算法的并行寻优特征和挥发系
数方法的自适应更改信息量的能力,并以球面聚类的方式确定了径向基神经网络中基函数的位置,
同时通过比较隐层神经元的相似性、合并相似性较为接近的2 个神经元来约简隐含层的神经元,以
达到简化径向基神经网络结构的目的. 实验比较了几种不同聚类算法的径向基神经网络,结果表
明,所提神经网络的整体训练时间至少可缩短40 ％ ,学习的准确率可提高1 ％以上,而且网络结构
更加精简.-An Ant clustering RBFNN presents a clustering algorithm based on the ant colony RBFN Liaison. Ant use the parallel algorithm optimization features and volatile coefficient method of adaptive changes in the volume of information, spherical cluster and the method of RBF neural network-based functions of the position, By comparing the same time hidden layer neurons of similarity, Merger similarity is closer to the two neurons to Jane about the hidden layer neurons, to achieve simplification RBF neural network structure purposes. Experimental comparison of several different clustering algorithm RBF neural network, The results show that the neural network's overall training time can be shortened at least 40%. Learning the accuracy rate can be increased by 1%, and the network is more streaml Platform: |
Size: 292558 |
Author:zhxj |
Hits:
Description: 经网络提出了一种基于蚁群聚类算法的径向基神经网络. 利用蚁群算法的并行寻优特征和挥发系数方法的自适应更改信息量的能力,并以球面聚类的方式确定了径向基神经网络中基函数的位置, 同时通过比较隐层神经元的相似性、合并相似性较为接近的2 个神经元来约简隐含层的神经元,以达到简化径向-via the Internet presents a clustering algorithm based on the ant colony RBFNN. Ant use the parallel algorithm optimization Volatile coefficient characteristics and the method of adaptive changes in the volume of information, spherical cluster and the method of RBF neural network-based functions of the position, By comparing the same time hidden layer neurons of similarity, Merger similarity is closer to the two neurons to Jane about the hidden layer neurons to achieve simplification Radial Platform: |
Size: 32313 |
Author:何泼 |
Hits:
Description: 一种基于蚁群聚类的径向基神经网络
提出了一种基于蚁群聚类算法的径向基神经网络. 利用蚁群算法的并行寻优特征和挥发系
数方法的自适应更改信息量的能力,并以球面聚类的方式确定了径向基神经网络中基函数的位置,
同时通过比较隐层神经元的相似性、合并相似性较为接近的2 个神经元来约简隐含层的神经元,以
达到简化径向基神经网络结构的目的. 实验比较了几种不同聚类算法的径向基神经网络,结果表
明,所提神经网络的整体训练时间至少可缩短40 ％ ,学习的准确率可提高1 ％以上,而且网络结构
更加精简.-An Ant clustering RBFNN presents a clustering algorithm based on the ant colony RBFN Liaison. Ant use the parallel algorithm optimization features and volatile coefficient method of adaptive changes in the volume of information, spherical cluster and the method of RBF neural network-based functions of the position, By comparing the same time hidden layer neurons of similarity, Merger similarity is closer to the two neurons to Jane about the hidden layer neurons, to achieve simplification RBF neural network structure purposes. Experimental comparison of several different clustering algorithm RBF neural network, The results show that the neural network's overall training time can be shortened at least 40%. Learning the accuracy rate can be increased by 1%, and the network is more streaml Platform: |
Size: 291840 |
Author:zhxj |
Hits:
Description: 经网络提出了一种基于蚁群聚类算法的径向基神经网络. 利用蚁群算法的并行寻优特征和挥发系数方法的自适应更改信息量的能力,并以球面聚类的方式确定了径向基神经网络中基函数的位置, 同时通过比较隐层神经元的相似性、合并相似性较为接近的2 个神经元来约简隐含层的神经元,以达到简化径向-via the Internet presents a clustering algorithm based on the ant colony RBFNN. Ant use the parallel algorithm optimization Volatile coefficient characteristics and the method of adaptive changes in the volume of information, spherical cluster and the method of RBF neural network-based functions of the position, By comparing the same time hidden layer neurons of similarity, Merger similarity is closer to the two neurons to Jane about the hidden layer neurons to achieve simplification Radial Platform: |
Size: 31744 |
Author:何泼 |
Hits:
Description: 实现了从wave中读取数据，并通过mfcc提取特征系数。再通过分裂法聚类，最后使用EM算法建立GMM。-Realize the data read from the wave, and through MFCC feature extraction coefficient. Clustering through secession law, and finally the use of EM algorithm for the establishment of GMM. Platform: |
Size: 5120 |
Author:gujunjun |
Hits:
Description: 基于matlab的关于复杂网络中的群聚系数的算法源程序，与大家共享-Matlab based on the complexity of the network clustering coefficient algorithm source code, and share Platform: |
Size: 1024 |
Author:huanglimin |
Hits:
Description: 四种聚类算法源代码及示例代码,本程序的最终目的是形成一套标准的用于聚类、可扩展的工具。包括的内容有1. 聚类算法：Kmeans和Kmedoid算法、FCMclust, GKclust, GGclust算法 2. 评估分类原型：程序可以在二维图像上绘制出聚类的结果 3. 验证：程序给每一个算法提供验证机制，每个聚类算法会统计Partition Coefficient (PC), Classification Entropy (CE), Partition Index (SC), Separation Index (S), Xie and Beni s Index (XB), Dunn s Index (DI) and Alternative Dunn Index (DII)几种衡量指标。-Four clustering algorithm source code and sample code, the ultimate goal of this process is the formation of a set of criteria for clustering, scalable tool. Including the contents of 1. Clustering Algorithm: Kmeans and Kmedoid algorithm, FCMclust, GKclust, GGclust algorithm 2. Assess the classification of a prototype: The procedure can be two-dimensional images to map out the results of clustering 3. Verify: Program to provide every algorithm authentication mechanism, each clustering algorithm statistics Partition Coefficient (PC), Classification Entropy (CE), Partition Index (SC), Separation Index (S), Xie and Beni s Index (XB), Dunn s Index (DI) and Alternative Dunn Index (DII) of several indicators. Platform: |
Size: 2136064 |
Author:王梦萍 |
Hits:
Description: 程序包包含的验证方法，会根据以下的多个指数分配系数(Partition Coefficient)，分类熵Classification Entropy ，分区索引，分离指数（Separation Index），Xie and Beni s的索引，严重聚类的算法-Package contains the authentication method will be based on the following number of index coefficient (Partition Coefficient), classification entropy Classification Entropy, geographical index, separation index (Separation Index), Xie and Beni s index, a serious clustering algorithm Platform: |
Size: 2066432 |
Author:xuechaoling |
Hits:
Description: 四种聚类算法源代码及示例代码,本程序的最终目的是形成一套标准的用于聚类、可扩展的工具。包括的内容有1. 聚类算法：Kmeans和Kmedoid算法、FCMclust, GKclust, GGclust算法 2. 评估分类原型：程序可以在二维图像上绘制出聚类的结果 3. 验证：程序给每一个算法提供验证机制，每个聚类算法会统计Partition Coefficient (PC), Classification Entropy (CE), Partition Index (SC), Separation Index (S), Xie and Beni s Index (XB), Dunn s Index (DI) and Alternative Dunn Index (DII)几种衡量指标。-Four kinds of clustering algorithms source code and sample code, the ultimate goal of this program is to form a set of standards for clustering, scalable tools. Including the contents of a. Clustering algorithm: Kmeans and Kmedoid algorithms, FCMclust, GKclust, GGclust algorithm 2. Assessment classified prototypes: programs can be plotted in two-dimensional image, the results of clustering 3. Verify: Program to provide an algorithm to each authentication mechanism, each clustering algorithm will calculate Partition Coefficient (PC), Classification Entropy (CE), Partition Index (SC), Separation Index (S), Xie and Beni' s Index (XB), Dunn' s Index (DI) and Alternative Dunn Index (DII) of several metrics. Platform: |
Size: 2134016 |
Author:conroy cheung |
Hits:
Description: 本程序的最终目的是形成一套标准的用于聚类、可扩展的工具。包括的内容有1. 聚类算法：Kmeans和Kmedoid算法、FCMclust, GKclust, GGclust算法 2. 评估分类原型：程序可以在二维图像上绘制出聚类的结果 3. 验证：程序给每一个算法提供验证机制，每个聚类算法会统计Partition Coefficient (PC), Classification Entropy (CE), Partition Index (SC), Separation Index (S), Xie and Beni s Index (XB), Dunn s Index (DI) and Alternative Dunn Index (DII)几种衡量指标。-The ultimate goal of this program is to form a set of standards for clustering, scalable tools. Including the contents of 1 clustering algorithm: Kmeans and Kmedoid algorithm, FCMclust, GKclust, GGclust algorithm 2. Assessment classified prototypes: the program can draw two-dimensional image on the results of clustering 3 Verify: Program to provide an algorithm to each authentication mechanism, the statistics of each clustering algorithm Partition Coefficient (PC), Classification Entropy (CE), Partition Index (SC), Separation Index (S), Xie and Beni' s Index (XB), Dunn' s Index (DI) and Alternative Dunn Index (DII) of several metrics. Platform: |
Size: 2141184 |
Author:xcm |
Hits:
Description: 1998年, Watts和Strogatz 提出了小世界网络这一概念,并建立了WS模型。 实证结果表明,大多数的真实网络都具有小世界特性(较小的最短路径) 和聚类特性(较大的聚类系数) 。
WS小世界模型构造算法
1、从规则图开始：考虑一个含有N个点的最近邻耦合网络，它们围成一个环，其中每个节点都与它左右相邻的各K/2节点相连，K是偶数。
2、随机化重连：以概率p随机地从新连接网络中的每个边，即将边的一个端点保持不变，而另一个端点取为网络中随机选择的一个节点。其中规定，任意两个不同的节点之间至多只能有一条边，并且每一个节点都不能有边与自身相连。
-In 1998, Watts and Strogatz proposed the concept of the small world network, and the establishment of the WS model. The empirical results show that most real networks have small world (the smaller the shortest path) and clustering properties (high clustering coefficient).
WS small-world model construction algorithm
1, from the rules FIG: consider a nearest neighbor coupling network with N points, they are surrounded by a ring, wherein each node and its left and right adjacent to each of K/2 connected to the node, K is an even number.
The 2 randomized reconnection: randomly with probability p from the new connection on each side of the network, coming section of one of the endpoints remain unchanged, while the other endpoint is taken as the random selection of a node in the network. Which states that between any two different nodes can have at most one edge, and each node can not be connected to its own side. Platform: |
Size: 1140736 |
Author:李梅 |
Hits:
Description: 四种聚类算法源代码及示例代码,本程序的最终目的是形成一套标准的用于聚类、可扩展的工具。包括的内容有1. 聚类算法：Kmeans和Kmedoid算法、FCMclust, GKclust, GGclust算法 2. 评估分类原型：程序可以在二维图像上绘制出聚类的结果 3. 验证：程序给每一个算法提供验证机制，每个聚类算法会统计-Partition Coefficient (PC), classification Entropy (CE), Partition Index (SC), Separation Index (S), Xie and Beni s Index (XB), Dunn s Index (DI) and Alternative Dunn Index (DII)几种衡量指标。-Four kinds of clustering algorithms source code and sample code, the ultimate goal of this program is to form a set of standards for clustering, scalable tools. Including the contents of a. Clustering algorithm: Kmeans and Kmedoid algorithms, FCMclust, GKclust, GGclust algorithm 2. Assessment classified prototypes: programs can be plotted in two-dimensional image, the results of clustering 3. Verify: Program to provide an algorithm to each authentication mechanism, each clustering algorithm will calculate Partition Coefficient (PC), classification Entropy (CE), Partition Index (SC), Separation Index (S), Xie and Beni s Index (XB), Dunn s Index (DI) and Alternative Dunn Index (DII) of several metrics. Platform: |
Size: 2073600 |
Author:chendi |
Hits:
Description: 本算法能够按照两种策略生成并绘制随机图，但是对于生成的随机图的属性分析不够完善，从代码中能够看出作者并没有完成本算法关于“平均路径长度”、“聚类系数”和“平均度”三个属性的求解。-The algorithm is able to follow two strategies to generate and draw random graph, but for the properties generated random graph analysis is not perfect, it can be seen from the code of the algorithm is not complete on the "average path length", "clustering coefficient" and "the average degree of" solving the three properties Platform: |
Size: 1024 |
Author:薇薇 |
Hits:
Description: 本算法实现了BA无标度模型的构造过程，并提供了聚类系数等算法。-The algorithm of the construction process of BA scale models and provides clustering coefficient algorithm. Platform: |
Size: 12288 |
Author:李见 |
Hits:
Description: 用于时频分析算法，可实现对二维数据的聚类，脉冲响应的相关分析算法并检验，最终的权值矩阵就是滤波器的系数，基于分段非线性权重值的Pso算法，插值与拟合，解方程，数据分析。- For time-frequency analysis algorithm, Can realize the two-dimensional data clustering, Related impulse response analysis algorithm and inspection, The final weight matrix is ??the filter coefficient, Based on piecewise nonlinear weight value Pso algorithm, Interpolation and fitting, solution of equations, data analysis. Platform: |
Size: 9216 |
Author:zunvh |
Hits:
Description: 旋转机械二维全息谱计算，最终的权值矩阵就是滤波器的系数，有小波分析的盲信号处理，详细画出了时域和频域的相关图，主同步信号PSS在时域上的相关仿真，插值与拟合，解方程，数据分析，ICA（主分量分析）算法和程序，基于欧几里得距离的聚类分析。- Rotating machinery 2-d holographic spectrum calculation, The final weight matrix is ??the filter coefficient, There Wavelet Analysis Blind Signal Processing, Correlation diagram shown in detail the time domain and frequency domain, PSS primary synchronization signal in the time domain simulation related, Interpolation and fitting, solution of equations, data analysis, ICA (Principal Component Analysis) algorithm and procedures, Clustering analysis based on Euclidean distance. Platform: |
Size: 6144 |
Author:vepni |
Hits:
Description: 预报误差法参数辨识-松弛的思想，基于kaiser窗的双谱线插值FFT谐波分析，独立成分分析算法降低原始数据噪声，MIMO OFDM matlab仿真,感应双馈发电机系统的仿真，最终的权值矩阵就是滤波器的系数，基于欧几里得距离的聚类分析，包括随机梯度算法，相对梯度算法。- Prediction Error Method for Parameter Identification- the idea of relaxation, Dual-line interpolation FFT harmonic analysis kaiser windows, Independent component analysis algorithm reduces the raw data noise, MIMO OFDM matlab simulation, Simulation of doubly fed induction generator system, The final weight matrix is ??the filter coefficient, Clustering analysis based on Euclidean distance, Including stochastic gradient algorithm, the relative gradient algorithm. Platform: |
Size: 7168 |
Author:exgpsyzwp |
Hits:
Description: 一种面向工艺规划的零件分类和识别方法：采用零件工艺相似系数对零件实现分类。类似K-means的聚类算法。-Part classification and recognition method for process planning: the classification of parts by the process similarity coefficient. A clustering algorithm similar to K-means. Platform: |
Size: 1024 |
Author:蒋林 |
Hits:
Search - clustering coefficient algorithm