Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/MingyanZHU/machine_learning
/ functions
Functions
68 in github.com/MingyanZHU/machine_learning
⨍
Functions
68
◇
Types & classes
12
↓ 10 callers
Function
predict
用于拟合函数 Args: X_Test 为 (m+1, m+1)的矩阵 w 为求解得到的系数向量,其维度为m
lab1/lab1.py:47
↓ 5 callers
Function
accuracy
计算在给定测试集上的分类准确度
lab2/lab2.py:108
↓ 5 callers
Method
fitting
(self)
lab2/newton_method.py:32
↓ 4 callers
Method
fitting
用于牛顿法求解方程的解
lab1/newton_method.py:31
↓ 3 callers
Method
__euclidean_distance
(x1, x2)
lab3/k_means.py:16
↓ 3 callers
Function
pca
进行PCA(Principal Component Analysis) data:为原始数据 reduced_dimension:为需要降低到的维数
lab4/lab4.py:58
↓ 3 callers
Function
transform
Transform an array to (len(X), degree + 1) matrix. Args: X: an ndarray. degree:int, degree for polynomial. Returns:
lab1/lab1.py:27
↓ 2 callers
Method
E_rms
根均方(RMS)误差
lab1/analytical_solution.py:19
↓ 2 callers
Method
__expectation
(self)
lab3/gaussian_mixture_model.py:65
↓ 2 callers
Method
__initial_center_not_random
选择彼此距离尽可能远的K个点
lab3/k_means.py:19
↓ 2 callers
Method
__k_means
(self)
lab3/k_means.py:33
↓ 2 callers
Method
__loss
(self, beta_t)
lab2/gradient_descent.py:18
↓ 2 callers
Method
__loss
用于求解loss 在本次实验中loss的公式为 $E(w) = \frac{1}{2N}[(Xw - T)'(Xw - T) + \lambda w'w]$
lab1/gradient_descent.py:18
↓ 2 callers
Method
__sigmod
(self, z)
lab2/newton_method.py:14
↓ 2 callers
Method
__sigmod
(self, z)
lab2/gradient_descent.py:15
↓ 2 callers
Method
acc
用于测试聚类的正确率
lab3/iris_read.py:17
↓ 2 callers
Function
draw_2_dimensions
画出二维参数的样本
lab2/lab2.py:120
↓ 2 callers
Method
fitting_with_regulation
带惩罚项的解析解
lab1/analytical_solution.py:14
↓ 2 callers
Method
get_data
(self)
lab2/blood_read.py:10
↓ 2 callers
Method
k_means_not_random_center
随机选择第一个簇中心点 再选择彼此距离最大的k个顶点作为初始簇中心点
lab3/k_means.py:64
↓ 2 callers
Method
predict
(self)
lab3/gaussian_mixture_model.py:97
↓ 2 callers
Function
psnr
计算信噪比
lab4/lab4.py:74
↓ 2 callers
Function
sigmod
(z)
lab2/lab2.py:10
↓ 2 callers
Function
split_data
分开训练集与测试集 Args: test_rate 为默认测试集占所有样本的比例
lab2/lab2.py:72
↓ 1 callers
Method
__converged
(self)
lab3/gaussian_mixture_model.py:84
↓ 1 callers
Method
__derivative
(self, beta_t)
lab2/newton_method.py:17
↓ 1 callers
Method
__derivative
求函数的一阶导数 即 $J(w) = (X'X + \lambda I)w - X'T$
lab1/newton_method.py:18
↓ 1 callers
Method
__derivative
一阶函数求导
lab1/gradient_descent.py:27
↓ 1 callers
Method
__derivative_beta
(self, beta_t)
lab2/gradient_descent.py:25
↓ 1 callers
Method
__euclidean_distance
(x1, x2)
lab3/gaussian_mixture_model.py:26
↓ 1 callers
Method
__init_params
(self)
lab3/gaussian_mixture_model.py:43
↓ 1 callers
Method
__initial_center_not_random
选择彼此距离尽可能远的K个点
lab3/gaussian_mixture_model.py:29
↓ 1 callers
Method
__likelihoods
(self)
lab3/gaussian_mixture_model.py:59
↓ 1 callers
Method
__maximization
(self)
lab3/gaussian_mixture_model.py:75
↓ 1 callers
Method
__second_derivative
求二阶导黑塞矩阵的逆
lab2/newton_method.py:24
↓ 1 callers
Method
__second_derivative
求函数的二阶导数的逆 即hessian矩阵的逆
lab1/newton_method.py:25
↓ 1 callers
Function
draw_data
将PCA前后的数据进行可视化对比
lab4/lab4.py:41
↓ 1 callers
Function
generateData
Generate training or test data. Args: number: data number you want which is an integer scale: the variance of Gaussian diribution
lab1/lab1.py:9
↓ 1 callers
Function
generate_2_dimension_data
当条件独立的时候，以cov11 cov12 cov21 cov22构成的协方差阵为对角阵 1.当不满足sigma仅与类相关的条件时候，即sigma与维度和类均相关，可令scale_pos1_bios与scale_pos2_bios不相同(同理使 scale_neg1_bi
lab2/lab2.py:14
↓ 1 callers
Function
generate_data
生成2维或者3维数据
lab4/lab4.py:12
↓ 1 callers
Function
generate_data
生成2维数据 :argument number_k k类 :argument sample_means k类数据的均值 以list的形式给出 如[[1, 2],[-1, -2], [0, 0]] :argument sample_number k类数据的数量 以list的形
lab3/lab3.py:43
↓ 1 callers
Method
get_data
(self)
lab3/iris_read.py:14
↓ 1 callers
Method
k_means_random_center
随机选择k个顶点作为初始簇中心点
lab3/k_means.py:59
↓ 1 callers
Function
load_mnist
读取mnist中的训练数据
lab4/lab4.py:28
↓ 1 callers
Function
main
()
lab2/blood_read.py:12
Method
__init__
(self)
lab2/mushroom_read.py:7
Method
__init__
(self)
lab2/blood_read.py:6
Method
__init__
(self, x, y, beta_0, hyper=0, delta=1e-6)
lab2/newton_method.py:5
Method
__init__
(self)
lab2/bank_note_read.py:7
Method
__init__
(self, x, y, beta_0, hyper=0, rate=0.1, delta=1e-6)
lab2/gradient_descent.py:5
Method
__init__
共轭梯度下降 Args: X, T 训练集 其中X为(m+1, m+1)的矩阵 T为(N, 1)的向量 hyper 超参数 delta 迭代停止条件
lab1/conjugate_gradient.py:5
Method
__init__
求方程的解析解
lab1/analytical_solution.py:5
Method
__init__
Args: X, T训练集, 其中X为(number_train, degree + 1)的矩阵 T为(number_train, 1)的向量 hyper 为超参数 delta 为迭代
lab1/newton_method.py:5
Method
__init__
Args: X, T训练集, 其中X为(number_train, degree + 1)的矩阵 T为(number_train, 1)的向量 hyper 为超参数 rate 为学习率, delta 为
lab1/gradient_descent.py:5
Method
__init__
(self, data, k=3, delta=1e-12, max_iteration=1000)
lab3/gaussian_mixture_model.py:10
Method
__init__
(self)
lab3/iris_read.py:6
Method
__init__
(self, data, k, delta=1e-6)
lab3/k_means.py:7
Method
fitting
(self)
lab2/gradient_descent.py:32
Method
fitting
(self, w_0)
lab1/conjugate_gradient.py:20
Method
fitting
不带惩罚项的解析解
lab1/analytical_solution.py:10
Method
fitting
用于多项式函数使用梯度下降拟合 Args: w_0 初始解，通常以全零向量
lab1/gradient_descent.py:33
Method
fitting_standford
(self, w_0)
lab1/conjugate_gradient.py:38
Method
get_data
(self)
lab2/mushroom_read.py:20
Method
get_data
(self)
lab2/bank_note_read.py:12
Function
load_watermelon_data
读取西瓜书3.3a的数据
lab2/lab2.py:100
Method
predict
(self, beta)
lab2/gradient_descent.py:53
Function
psnr_trace
计算全部的信噪比
lab4/lab4.py:82
Function
watermelon_data
()
lab3/lab3.py:9