Coursera机器学习课程笔记(五)
由于好几题都是在nnCostFunction.m这一个文件中完成,所以一并给出1、2、4和5的代码
1.Feedforward and Cost Function
2.Regularized Cost Function
4.Neural Network Gradient (Backpropagation)
5.Regularized Gradient
nnCostFunction.m文件:
function [J grad] = nnCostFunction(nn_params, ...
input_layer_size, ...
hidden_layer_size, ...
num_labels, ...
X, y, lambda)
%NNCOSTFUNCTION Implements the neural network cost function for a two layer
%neural network which performs classification
% [J grad] = NNCOSTFUNCTON(nn_params, hidden_layer_size, num_labels, ...
% X, y, lambda) computes the cost and gradient of the neural network. The
% parameters for the neural network are "unrolled" into the vector
% nn_params and need to be converted back into the weight matrices.
%
% The returned parameter grad should be a "unrolled" vector of the
% partial derivatives of the neural network.
%
% Reshape nn_params back into the parameters Theta1 and Theta2, the weight matrices
% for our 2 layer neural network
Theta1 = reshape(nn_params(1:hidden_layer_size * (input_layer_size + 1)), ...
hidden_layer_size, (input_layer_size + 1));
Theta2 = reshape(nn_params((1 + (hidden_layer_size * (input_layer_size + 1))):end), ...
num_labels, (hidden_layer_size + 1));
% Setup some useful variables
m = size(X, 1);
% You need to return the following variables correctly
J = 0;
Theta1_grad = zeros(size(Theta1));
Theta2_grad = zeros(size(Theta2));
% ====================== YOUR CODE HERE ======================
% Instructions: You should complete the code by working through the
% following parts.
%
% Part 1: Feedforward the neural network and return the cost in the
% variable J. After implementing Part 1, you can verify that your
% cost function computation is correct by verifying the cost
% computed in ex4.m
%
% Part 2: Implement the backpropagation algorithm to compute the gradients
% Theta1_grad and Theta2_grad. You should return the partial derivatives of
% the cost function with respect to Theta1 and Theta2 in Theta1_grad and
% Theta2_grad, respectively. After implementing Part 2, you can check
% that your implementation is correct by running checkNNGradients
%
% Note: The vector y passed into the function is a vector of labels
% containing values from 1..K. You need to map this vector into a
% binary vector of 1's and 0's to be used with the neural network
% cost function.
%
% Hint: We recommend implementing backpropagation using a for-loop
% over the training examples if you are implementing it for the
% first time.
%
% Part 3: Implement regularization with the cost function and gradients.
%
% Hint: You can implement this around the code for
% backpropagation. That is, you can compute the gradients for
% the regularization separately and then add them to Theta1_grad
% and Theta2_grad from Part 2.
%
%模拟神经网络的前向传播过程,Theta1和Theta2是已经训练好的
a1 = X; %第一层
a1 = [ones(m,1),a1];
z2 = a1 * Theta1';
a2 = sigmoid(z2); %第二层
a2 = [ones(m,1),a2];
z3 = a2 * Theta2';
a3 = sigmoid(z3); %第三层,a3可看作是输出结果h(x)
for k = 1:num_labels
%根据公式,计算当k为不同数字时的误差之和,通过a3(:,k)来取出对应的列与(y==k)矩阵进行运算
J = J + (1/m) * (-(y == k)' * log(a3(:,k)) - (1-(y == k))' * log(1 - a3(:,k)));
end
%下面单独计算正则化的尾项
sum = 0;
for i = 1:hidden_layer_size
for j = 2:input_layer_size + 1 % Theta1第一列不计算
sum = sum + Theta1(i,j)^2;
end
end
for i = 1:num_labels
for j = 2:hidden_layer_size + 1 %Theta2第一列不计算
sum = sum + Theta2(i,j)^2;
end
end
J = J + lambda/(2*m) * sum; %包含正则化的损失函数J
% -------------------------------------------------------------
%反向传播算法
delta_1 = 0;
delta_2 = 0;
for t = 1:m %根据要求,分成m次循环实现反向传播,m代表训练样本的个数
d_3 = a3(t,:)' - ([1:num_labels] == y(t,1))'; %对应step2,d_3为10x1矩阵
d_2 = Theta2(:,2:end)' * d_3 .* sigmoidGradient(z2(t,:)'); %对应step3,注意Theta2的第一列为bias所以舍去,d_2为25x1矩阵
delta_1 = delta_1 + d_2 * a1(t,:); %对应step4
delta_2 = delta_2 + d_3 * a2(t,:); %对应step4
end
%包含正则化的梯度
Theta1(:,1) = 0; %把第一列全置0,不需要考虑正则化
Theta1_grad = delta_1 / m + lambda / m * Theta1;
Theta2(:,1) = 0;
Theta2_grad = delta_2 / m + lambda / m * Theta2;
% ========================================================================
% Unroll gradients
grad = [Theta1_grad(:) ; Theta2_grad(:)];
end
关于反向传播算法这一部分的代码,强烈建议在for循环处建立断点,在调试过程观察各个矩阵的维度,就能理解代码在做什么了。
3.Sigmoid Gradient
sigmoidGradient.m文件:
function g = sigmoidGradient(z)
%SIGMOIDGRADIENT returns the gradient of the sigmoid function
%evaluated at z
% g = SIGMOIDGRADIENT(z) computes the gradient of the sigmoid function
% evaluated at z. This should work regardless if z is a matrix or a
% vector. In particular, if z is a vector or matrix, you should return
% the gradient for each element.
g = zeros(size(z));
% ====================== YOUR CODE HERE ======================
% Instructions: Compute the gradient of the sigmoid function evaluated at
% each value of z (z can be a matrix, vector or scalar).
g = exp(-z) ./ (1 + exp(-z)).^2;
% =============================================================
end