weixin_43310853的博客

文章目录Error analysisMethods to solve over fittingMethods to solve under fittingRecommend approachError metrics for skewed classesData for machine learningSupport Vector MachineK-meansPrinciple Component AnalysisData preprocessingChoosing the number of princi

文章目录1. 创建array2. 基础运算3. 索引4. array合并5. array分割6. copy和deepcopy import numpy as np 1. 创建array # 创建array,dtype格式 array = np.array( [[1, 2, 3], [2, 3, 4]], dtype=np.float) # shape print(array.shape) # 维数 print(array.ndim) # 总数量 print(array.size) (

Build a neural network from Scratch 文章目录Build a neural network from Scratch1. Node2. Variable3. Linear4. Relu5. Sigmoid6. MSE7. Session8. main 1. Node class Node: ''' name 该节点标识 inputs 指向该节点的节点 outputs 出节点 is_trainable 是

文章目录1.建立计算图2.拓扑排序3.前馈计算4.反向传播 1.建立计算图 定义变量和使用各种算子的过程就是一个建立一个计算图的过程 import networkx as nx import random import numpy from matplotlib.animation import FuncAnimation %matplotlib notebook seed=1 random.seed(seed) np.random.seed(seed) sample_graph = { 'x'