搜索资源列表
fault-programm
- 对变压器的故障识别模式进行了编程,可以很清楚的给出各故障发生趋势。-This is a program that it could recognize the faults of transformer and it could predict the faults.
orc1
- 本程序采用svm分类算法,实现了对连体数字(0-100)的识别。 分两个过程,第一个需要训练生成mode,第二个过程根据mode进行预测。识别率达95 -This program uses the svm classification algorithm, the identification of the piece number (0-100). Divided into two processes, first need to be trained to generate mode, th
Naive-Bayes
- Naive Bayes classifier Input: Training set and testing set, each row represents a instance, last column is label, begins from zero Output:predict label by Naive Bayes as well as its accuracy
N-Gram-LM
- 该程序基于Bi-Gram模型算法思想对一训练文本中的词汇建立语言模型,而后对测试文本中的语句进行预测出现概率-The program is based on Bi-Gram model algorithm ideas on a training text vocabulary build language model, and then the text of the statement of the test to predict the probability of occurrence
LVQ-NATURE-NETWORK-
- 基于MATLAB的SVQ人脸朝向识别,所有代码和识别图片均在压缩文件中-LVQ neural network to predict- face toward the identification
677596
- 本程序根据训练好的网络文件ANN,mat预测新的数据文件 很好的(This procedure according to the trained network file ANN, mat predict new data file is very good)
keras_mnist_test
- hello Word of keras ,第一个成功实现的卷积神经网络,下载了mnist数据集,并decode,,然后,为了加快速度,训练其中的一部分数据,并用predict测试,ok,2min内出结果.(网上其它程序试过,训练太久,一晚上都没训练出结果,于是自己动手设计了这个小程序) 环境:python3.6,keras2.1,PC i5(很破的电脑)(Hello Word of keras, the first successful convolution neural network,
UCI的光学字符识别数据集
- 其目标是将大量黑白矩形像素显示器中的每一个识别为英文字母中的26个大写字母之一。字符图像基于20种不同的字体,并且这20种字体中的每个字母随机失真以产生20,000个独特刺激的文件。每个刺激被转换成16个基本的数字属性(统计矩和边缘计数),然后将其缩放以适合从0到15的整数值范围。我们通常在前16000个项目上进行训练,然后使用结果模型预测剩余的4000个字母类别。请参阅上面引用的文章以获取更多详细信息。(The objective is to identify each of a large