搜索资源列表
MFL
- 基于改进BP神经网络算法的管道缺陷漏磁信号识别-Identification of magnetic flux leakage signals of Pipeline Defects Based on improved BP neural network algorithm
jh447
- 包括邓氏关联度、绝对关联度、斜率关联度、改进绝对关联度,一个师兄的毕设,BP神经网络的整个训练过程。- Including Deng s correlation, absolute correlation, correlation of slope, improved absolute correlation, A complete set of brothers, The entire training process BP neural network.
gannuiqai
- 有信道编码,调制,信道估计等,BP神经网络的整个训练过程,包括邓氏关联度、绝对关联度、斜率关联度、改进绝对关联度。- Channel coding, modulation, channel estimation, The entire training process BP neural network, Including Deng s correlation, absolute correlation, correlation of slope, improved absolute corr
Image-filter.tar
- 提出了一种基于改进 BP 神经网络和粒子群优化算法( PSO) 的图像滤波方法 。该方法利用双曲正切形式 的误差函数代替 BP 神经网络传统的最小均方误差函数( LMS),并将改进后的 BP 神经网络利用 PSO 算法优 化,用来减小图像噪声对神经网络精度的影响以及避免神经网络陷入局部极小值点,从而提高神经网络去噪能 力。实验结果表明,与传统滤波方法相比,该方法不仅能有效地滤除图像中的高斯噪声而且能很好地保护图像 细节 。- U63D0 u51FA u4E86 u4E00
Training
- 极限学习机的回归拟合——对比实验研究,改进BP算法的一些缺点,在ELM算法的基础上进行回归拟合预测。(Regression fitting Some shortcomings of BP algorithm are improved, and regression fitting prediction is carried out on the basis of ELM algorithm.)
fnel
- bp算法及改进bp算法,已测试,附带说明文档()
BYEHLY
- bp算法及改进bp算法,已测试,附带说明文档()
26个字母识别 用matlab实现的
- 用matlab实现的26个字母识别。基于BP算法的字母识别其容错性和识别率相对较高,在有噪声的情况下训练其识别出错率也相应增加,许进一步改进。(26 letter recognition implemented with MATLAB.The letter recognition based on BP algorithm has a relatively high fault tolerance and recognition rate, and the error rate of recog
神经网络
- 用matlab编程进行神经网络算法的运用。内部含有SVM及SVM的改进,PSO算法,BP算法等等。(Matlab programming for the use of neural network algorithm. Internal SVM and SVM improvement, PSO algorithm, BP algorithm and so on.)