ex1练习

最新推荐文章于 2022-10-31 18:55:38 发布

theeb

最新推荐文章于 2022-10-31 18:55:38 发布

阅读量260

点赞数

本文链接：https://blog.csdn.net/xuyuehengyi/article/details/95237874

版权

加载数据

% 单变量
data = load('ex1data1.txt');
X = data(:, 1);   % m*1矩阵
y = data(:, 2);   % m*1矩阵
m = length(y); % number of training examples

% 多变量
data = load('ex1data2.txt');
X = data(:, 1:2); % m*2矩阵
y = data(:, 3);   % m*1矩阵
m = length(y);

代价函数（两者公式一样，多变量为例）计算不同theta下的代价

X = [ones(m, 1) X];
theta = zeros(3, 1);
h_theta = X*theta; % 假设函数，X是m*n矩阵，theta是n*1列向量，m样本数，n特征数
J = (1/(2*m))*(h_theta-y)'*(h_theta-y);

$h_\theta(x)=\theta^\intercal x=\theta_0+\theta_1x_1+\theta_2x_2$

$J(\theta)=\frac{1}{2m}\sum_{i=1}^{m}(h_\theta(x^i)-y^i)^2$

梯度下降更新theta值

alpha = 0.01;
num_iters = 400;
for iter = 1:num_iters % 更新num_iters次
    h_theta = X*theta; % 假设函数
    theta = theta-(alpha/m)*((h_theta-y)'*X)'; 
        % ((h_theta-y)'*X)得到1*n向量，n是特征数，theta是n*1向量
end

$\theta_j=\theta_j-\frac{\alpha}{m}\sum_{i=1}^{m}(h_\theta(x^i)-y^i)x_j^i$

特征归一化（梯度下降提高收敛速度，正规方程不用）

mu = mean(X); % 对每一列求均值
sigma = std(X); % 对每一列求标准差
for i=1:size(X,1) % 对每一行进行归一化
	X_norm(i,:) = (X(i,:)-mu)./sigma;  
end

正规方程

theta = pinv(X'*X)*X'*y;

绘图
收敛曲线

figure;
plot(1:numel(J_history), J_history, '-b', 'LineWidth', 2); % numel返回J_history个数
xlabel('Number of iterations');
ylabel('Cost J');

绘制样本点和拟合曲线

% 绘制样本点
plot(x,y,'rx','MarkerSize',10);
ylabel('Profit in $10,000s');
xlabel('Population of City in 10,000s');
% 绘制假设函数曲线/拟合曲线
hold on; % keep previous plot visible
plot(X(:,2), X*theta, '-')
legend('Training data', 'Linear regression')
hold off

绘制不同theta下代价函数

theta0_vals = linspace(-10, 10, 100);
theta1_vals = linspace(-1, 4, 100);
J_vals = zeros(length(theta0_vals), length(theta1_vals));
% Fill out J_vals
for i = 1:length(theta0_vals)
    for j = 1:length(theta1_vals)
	  t = [theta0_vals(i); theta1_vals(j)];
	  J_vals(i,j) = computeCost(X, y, t);
    end
end

J_vals = J_vals'; % surf和contour命令需要转置
figure;
surf(theta0_vals, theta1_vals, J_vals)
xlabel('\theta_0'); ylabel('\theta_1');

绘制不同theta下代价函数等高线图

figure;
contour(theta0_vals, theta1_vals, J_vals, logspace(-2, 3, 20))
xlabel('\theta_0'); ylabel('\theta_1');
hold on;
plot(theta(1), theta(2), 'rx', 'MarkerSize', 10, 'LineWidth', 2);

梯度下降(公式一样) 用于更新theta值

alpha = 0.01;
num_iters = 400;
for iter = 1:num_iters
    h_theta = X*theta;
    theta = theta-(alpha/m)*((h_theta-y)'*X)';
end

theeb

关注

0
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
ex1练习

加载数据% 单变量data = load('ex1data1.txt');X = data(:, 1); % m*1矩阵y = data(:, 2); % m*1矩阵m = length(y); % number of training examples% 多变量data = load('ex1data2.txt');X = data(:, 1:2); % m*2矩阵...
复制链接

扫一扫