KRR
所属分类:人工智能/神经网络/深度学习
开发工具:matlab
文件大小:3KB
下载次数:15
上传日期:2020-05-07 15:56:49
上 传 者:
GHao
说明: 核岭回归算法
输入数据集(需要分开存放训练集和测试集)
利用4重交叉验证法调参
最后输出分类准确率
(Kernel ridge regression algorithm
Input data set (training set and test set need to be stored separately)
Parameter adjustment by 4-fold cross validation
Final output classification accuracy)
文件列表:
KRR\kernel_matrix.m (2236, 2020-04-04)
KRR\KRR_RBF.m (756, 2020-04-04)
KRR\oneVrestCoding.m (187, 2020-04-04)
KRR\oneVrestDecoding.m (283, 2020-04-04)
KRR\TEST_KRR.m (3123, 2020-04-28)
KRR (0, 2020-05-07)
# Ensemble Kernel Ridge Regression (KRR)
This set of codes implements Kernel Ridge Regression (KRR) algorithm (exact form) to classify data.
You should do some research on your own to understand the basic priciples on how KRR Works.
To get started, you should refer to 'TEST_KRR.m'.
#### Parameters
kernel: kernel (See below)
kernel_param: kernel parameters (See below)
lambda: Regularisation parameter (Needs to be positive float)
#### Kernels
On this set of codes, we implemented 'RBF_kernel','lin_kernel','poly_kernel','wav_kernel' and'Chi_square' kernels.
Details can be found in 'kernel_matrix.m'.
You can use any kernel.
kernel_param is different for different kernel used.
For example: for 'RBF_kernel', kernel_param is gamma (which is a positive float value).
#### Usage
The main function to train and test Kernel Ridge Regrtession is:
"""
[TestAcc,TrainingTime,TestingTime] = KRR_RBF(trainX,trainY,testX,testY,ModelParameters);
"""
Inputs:
trainX Training data
trainY Training labels
testX Testing data
testY Testing labels
ModelParameters Model parameters
Outputs:
TestAcc Testing Accuracy
TrainingTime Time taken for training
TestingTime Time taken for testing
近期下载者:
相关文件:
收藏者: