搜索资源列表
pso_train_NN
- PSO算法具有快速收敛而且有很强的跳出局部最优从而找到全局最优点的能力，故可以用它来训练优化神经网络，此程序主要研究这个方面。-PSO algorithm is fast and has a strong convergence of jumping out of the local optimal thus find the most advantages of the overall capacity, it can be used
bp+rbf
- 用hopfield写的训练0-9数字的识别!精底很高!效果很不错!!基本可以识别!-used to write the training hopfield 0-9 digit identification! Precision high end! Very good results! ! Basic can be identified.
NNNJ
- 用MATLAB编写的一种新的神经网络训练程序，可以按照要求训练网络。-MATLAB prepared a new neural network training program, we can follow the training requirements of the network.
MIMO_OFDMchannelestimation
- 东南大学的一篇硕士论文，做的是训练序列和导频插入的估计方法-Southeast University in a master's thesis, so is the training sequence and inserted Pilot estimation method
ofdm
- 基于训练序列的OFDM信道估计仿真程序，可直接运行-Training sequence based on the OFDM channel estimation simulation program can be run directly
svm
- svm的分类原理介绍，以及它的训练算法研究和应用-introduce the principle of SVM classification, as well as its training algorithm research and application
trainbp
- BP人工神经网络训练源码，采用三层网络结构，固定节点数目，可以设定学习速度和学习次数。-BP artificial neural network training source, using the three-tier network architecture, a fixed number of nodes, you can set the number of learning speed and learning.
ImageDigits
- 基于模板的图片识别方式，先训练后识别,非常好用-Template-based photo identification, first training to identify, very easy to use
vehiclerecognition
- 汽车牌照的定位和识别，用bp神经网络训练并识别
matlab
- 该算法在优化方面性能比较好，可用于函数极值，解方程，并且在前向神经网络权值和网络结构的训练也表现了不错的性能-The algorithm in the optimization performance is better, can be used in extreme value function, solve equations, and the former to the neural network weights and net
lpc10_10
- 剑桥大学语音分析代码-LPC线性预测编码源程序，包括训练与测试代码。-Speech Analysis, University of Cambridge code-LPC linear predictive coding the source program, including training and test code.
moshishibie
- 通过已知的先验概率，在输入训练序列之后，可以对数值进行判别是属于哪一类的数据。简单可以应用于模式识别-Through the known a priori probability, after the training sequence in the input can be carried out on the discriminant value belongs to which type of data. Simple and c
rbf
- 可以方便利用训练数据建立径向基函数神经网络并绘制输入-输出网格图。-Can easily make use of training data to establish radial basis function neural network and the mapping of input- output grid.
pca-svm
- 本程序用于对训练样本提取独立主元，作为样本特征，并送入SVM分类器中训练图像的预处理中不取对数，也无须做幅度归一，由ICA的应用条件决定的。预处理后的图像以向量的形式按行排列-err
ANewImageRetrievalMethodUsingObjectSemantics
- 为了降低图像高层语义与低层视觉特征之间的语义差异,本文以对象描述模型为基础,提出利用机器转换模型获取图像高层语义的方法。本方法首先利用图像分割技术对图像进行分割,然后利用机器学习的方法,得到训练样本集中高层语义与分割后低层视觉特征之间的先验概率关系 在查询的过程中,利用得到的先验概率模型计算与高层语义所对应的最大概率视觉低层特征,最后利用该低层特征进行检索,达到缩短高层语义与低层特征之间的语义差异的目的。在一个拥有5000 幅图像的图像
Kmean
- 利用数据挖掘知识和vb实现K-均值聚类分析，本程序提供了训练样本和测试样本， 经过训练正确率达70-76％-Knowledge and use of data mining vb achieve K-means clustering analysis, the procedures provided training samples and test samples, after training in the correct ra
ganzhiqi
- 模式识别分类器，本代码是感知器，训练数据在压缩包的fisher1.txt中-Pattern recognition classifier, this code is the perceptron, the training data in the compressed package in fisher1.txt
asm
- 基于ASM的人脸识别程序，包括训练部分。由Ghassan Hamarneh编写-ASM-based face recognition procedures, including the training component. Prepared by Ghassan Hamarneh
licensetagorientationandidentify
- 包括源代码和可执行文件,还给出了训练数据以及输出层权值_16段投影数字.-Including source code and executable file, also gives the training data and the output layer weights paragraph _16 digital projector.
HMM
- HMM隐马尔科夫模型源码，实现模型的训练-Hidden Markov Model HMM source, the realization of model training