斯坦福机器学习实验之1-单变量线性回归(Linear Regression with One Variable)

1.数据图形化(Plotting the Data)

在进行任务之前，将数据可视化对我们理解数据很有用。
1)载入数据
函数: load;length

data = load('ex1data1.txt'); % read comma separated data
X = data(:, 1); y = data(:, 2);
m = length(y); % number of training examples

2)可视化数据:plotData.m
函数：figure；plot；title；xlabel;ylabel

figure
plot(x,y,'Color','g','LineStyle','none','Marker','x','MarkerSize',10); % Plot the data
title('plotData');
xlabel('Population of City in 10,000s'); %Set the x-axis label
ylabel('Profit in $10,000s'); % Set the y-axis label

2.梯度下降(Gradient Descent)

1)准备参数
函数：ones;zeros;

X = [ones(m, 1), data(:,1)]; % Add a column of ones to x
theta = zeros(2, 1); % initialize fitting parameters
iterations = 1500;
alpha = 0.01;

2)计算代价函数(Computing the cost):computeCost.m
函数：sum
运算符： . ^

function J = computeCost(X,y,theta)
%myFun - Description
% Long description

% square every elements in matrix A : A .^2
m = length(y); % number of training examples
h=X*theta;
J=1/(2*m)*sum((h-y).^2);
end

3.梯度下降

gradientDescent.m

% for i=1:10,
%   v(i)=2^i;
% end;
function [theta,J_history] = gradientDescent(X,y,theta,alpha,iterations)
   m=length(y);
   theta_s=theta;
   for i=1:iterations, 
     theta(1)=theta(1)-alpha/m*sum(X*theta_s-y);
     theta(2)=theta(2)-alpha/m*sum((X*theta_s-y).*X(:,2));
     theta_s=theta;

     % Save the cost in every iteration
     J_history(iterations)=computeCost(X,y,theta);
   end
   J_history
end

说明：用theta_s保留计算结果；分别计算theta(1)和theta(2),原因是theta为两行向量，而X和y都是m行矩阵，维度不匹配。

4.调试

1)matlab向量编号是从1开始的；
2)运算时注意维度的匹配。