liangzai07的博客

原创 随机梯度下降

import matplotlib.pyplot as pltfrom mpl_toolkits.mplot3d import Axes3D #绘制3D坐标的函数 import numpy as npimport randomQ0=random.random()Q1=random.random()Q2=random.random()a=0.01E=1e-04x=np.array([[2104,3], [1600,3], [2400,3],

原创 二类LDA原理推导

原创 决策树 推导理解

原创 强化学习Actor Critic Method

import gym, osfrom itertools import countimport paddleimport paddle.nn as nnimport paddle.optimizer as optimimport paddle.nn.functional as Ffrom paddle.distribution import Categoricalprint(paddle.__version__)device=paddle.get_device()env=gym.mak

原创 手写数字识别模型

import paddlefrom paddle.nn import Linearimport paddle.nn.functional as Fimport osimport numpy as npimport matplotlib.pyplot as plttrain_dataset=paddle.vision.datasets.MNIST(mode='train')train_data0=np.array(train_dataset[0][0])train_label_0=np.a

原创 增加卷积神经网络进行图像分类的层数

经过学习发现，下一层的in_channels只需要和上一层的out_channels保持一致即可，并不是规定或计算得到的数值。代码如下：importpaddleimportpaddle.nn.functionalasFfrompaddle.vision.transformsimportToTensorimportnumpyasnpimportmatplotlib.pyplotasplttransform=ToTensor()...

原创 主成分分析 PCA

import numpy as npdata=np.array([[2.5,0.5,2.2,1.9,3.1,2.3,2.0,1.0,1.5,1.1], [2.4,0.7,2.9,2.2,3.0,2.7,1.6,1.1,1.6,0.9]])p=np.average(data, axis=1)xp=p[0]yp=p[1]# xp=0# yp=0# n=10# for i in range(0,n):# xp+=data[0,i]/n# yp+=

原创 HMM viterbi

import numpy as npstates=np.array([0,1]) #rainy--0 sunny--1obser=np.array([0,1,2]) # k walk-0 shop-1 clean-2start=np.array([0.6,0.4])transition_probability=np.array([[0.7,0.3], [0.4,0.6]])

原创 HMM隐马尔可夫模型推导

原创 改进 欧拉算法

import matplotlib.pyplot as pltimport numpy as npx0=0y0=1t=0.1yb=1x=np.zeros(100)y=np.zeros(100)X=np.zeros(100)Y=np.zeros(100)for i in range(1,100): y0=1.1*y0-0.2*x0/y0 x0=x0+t x1=x0+t k0=y0-2*x0/y0 yb=y0+t*k0 k1=yb-2*x1/y

原创 神经网络和反向传播算法

原创 欧拉公式Euler‘s formula

import matplotlib.pyplot as pltimport numpy as npx=0y=1X=np.zeros(100)Y=np.zeros(100)h=0.1for i in range(1,100): y=1.1*y-0.2*x/y x=x+h X[i]=x Y[i]=y print(x,y)plt.plot(X,Y)plt.scatter(X,Y)plt.show()0.1 1.10.2 1.19181818181

原创 贝叶斯公式 作业

原创 欧拉公式Euler‘s formula

import matplotlib.pyplot as pltimport numpy as npx=0y=1h=0.1for i in range(1,10): y=1.1*y-0.2*x/y x=x+h print(x,y)# plt.plot(x0,y0)# plt.scatter(x0,y0)# plt.show()0.1 1.10.2 1.1918181818181820.30000000000000004 1.2774378337147220.

原创 感知器算法 有图

import numpy as npimport randomw0=0w1=0w2=0n=0.01e=2.72E=1e-04e0=10e1=10e2=10x=np.array([[0,0,0], [0,1,0], [1,0,0], [1,1,1]])# t=np.array([[0],# [0],# [0],# [1]])for a in ra

原创 感知器算法

import numpy as npimport randomw0=0.2w1=0.2w2=0.2n=0.01e=float(2.72)#E=float(10*e^(-4))E=float(0.18)e0=10e1=10e2=10x=np.array([[0,0,0], [0,1,0], [1,0,0], [1,1,1]])# t=np.array([[0],# [0],#

原创 梯度下降算法

import matplotlib.pyplot as pltfrom mpl_toolkits.mplot3d import Axes3D #绘制3D坐标的函数 import numpy as npimport randomQ0=random.random()Q1=random.random()Q2=random.random()a=0.01#e=2.718#E=float(10*e^(-4))E=float(0.18)x=np.array([[2104,3],

原创 numpy

data=np.array([1,2,3])dataarray([1, 2, 3])1.数组拆分垂直拆分：numpy.vsplit(数组，份数)c=np.arange(1,13).reshape(6,2)carray([[ 1, 2], [ 3, 4], [ 5, 6], [ 7, 8], [ 9, 10], [11, 12]])np.vsplit(c,3)[array([[1, 2],