np.eye(C)[Y.reshape(-1)]
时间: 2023-09-28 16:09:11 浏览: 166
This code creates a one-hot encoded matrix from a given array Y.
np.eye(C) creates a CxC identity matrix, where C is the number of classes.
Y.reshape(-1) flattens the input array Y into a 1-dimensional array.
Finally, the flattened array Y is used as an index to select the corresponding row from the identity matrix, resulting in a one-hot encoded matrix.
相关问题
# 定义昂贵的函数 def expensive_func(t): return np.sum(t**2 - 10*np.cos(2*np.pi*t) + 10) # 定义高斯核函数 def gaussian_kernel(x, y, theta): return np.exp(-theta * cdist(x, y)**2) # 定义对数似然函数 def log_likelihood(params, x, y): theta, sigma = params k = gaussian_kernel(x, x, theta) + sigma**2 * np.eye(len(x)) try: L = np.linalg.cholesky(k) except np.linalg.LinAlgError: return -np.inf alpha = np.linalg.solve(L.T, np.linalg.solve(L, y)) return -0.5*y.T.dot(alpha) - np.sum(np.log(np.diag(L))) - 0.5*len(x)*np.log(2*np.pi) # 定义预测函数 def predict(x, y, x0, theta, sigma): k = gaussian_kernel(x, x, theta) + sigma**2 * np.eye(len(x)) k0 = gaussian_kernel(x, x0.reshape(1, -1), theta) k00 = gaussian_kernel(x0.reshape(1, -1), x0.reshape(1, -1), theta) try: L = np.linalg.cholesky(k) except np.linalg.LinAlgError: return np.nan, np.nan alpha = np.linalg.solve(L.T, np.linalg.solve(L, y)) mu = k0.T.dot(alpha) v = k00 - k0.T.dot(np.linalg.solve(L.T, np.linalg.solve(L, k0))) return mu, v # 生成随机数据 np.random.seed(666) X = np.random.uniform(-20, 20, size=(200, 10)) y = np.array([expensive_func(x) for x in X]) # 优化超参数 initial_params = [1, 1] bounds = [(1e-5, None), (1e-5, None)] res = minimize(lambda params: -log_likelihood(params, X, y), initial_params, bounds=bounds) theta, sigma = res.x # 在随机点上进行预测 x0 = np.random.uniform(-20, 20, size=(1, 10)) mu, v = predict(X, y, x0, theta, sigma) # 计算误差 exact_val = expensive_func(x0) error = (exact_val - mu)**2 print("预测误差:", error) print("预测方差:", v)注释一下
这段代码主要实现了使用高斯过程进行回归分析。其中定义了一个昂贵的函数 `expensive_func`,该函数实现了在给定输入的情况下进行昂贵计算的功能。然后定义了一个高斯核函数 `gaussian_kernel`,用于计算输入数据的协方差矩阵。接着定义了对数似然函数 `log_likelihood`,用于计算给定超参数情况下的对数似然值。最后定义了预测函数 `predict`,用于在给定超参数和输入数据的情况下进行预测。在代码中,使用了 `minimize` 函数对超参数进行优化,并使用随机数据进行了预测。最终输出预测误差和预测方差。
将这段代码转换为伪代码:def levenberg_marquardt(fun, grad, jacobian, x0, iterations, tol): """ Minimization of scalar function of one or more variables using the Levenberg-Marquardt algorithm. Parameters ---------- fun : function Objective function. grad : function Gradient function of objective function. jacobian :function function of objective function. x0 : numpy.array, size=9 Initial value of the parameters to be estimated. iterations : int Maximum iterations of optimization algorithms. tol : float Tolerance of optimization algorithms. Returns ------- xk : numpy.array, size=9 Parameters wstimated by optimization algorithms. fval : float Objective function value at xk. grad_val : float Gradient value of objective function at xk. grad_log : numpy.array The record of gradient of objective function of each iteration. """ fval = None # y的最小值 grad_val = None # 梯度的最后一次下降的值 x_log = [] # x的迭代值的数组,n*9,9个参数 y_log = [] # y的迭代值的数组,一维 grad_log = [] # 梯度下降的迭代值的数组 x0 = asarray(x0).flatten() if x0.ndim == 0: x0.shape = (1,) # iterations = len(x0) * 200 k = 1 xk = x0 updateJ = 1 lamda = 0.01 old_fval = fun(x0) gfk = grad(x0) gnorm = np.amax(np.abs(gfk)) J = [None] H = [None] while (gnorm > tol) and (k < iterations): if updateJ == 1: x_log = np.append(x_log, xk.T) yk = fun(xk) y_log = np.append(y_log, yk) J = jacobian(x0) H = np.dot(J.T, J) H_lm = H + (lamda * np.eye(9)) gfk = grad(xk) pk = - np.linalg.inv(H_lm).dot(gfk) pk = pk.A.reshape(1, -1)[0] # 二维变一维 xk1 = xk + pk fval = fun(xk1) if fval < old_fval: lamda = lamda / 10 xk = xk1 old_fval = fval updateJ = 1 else: updateJ = 0 lamda = lamda * 10 gnorm = np.amax(np.abs(gfk)) k = k + 1 grad_log = np.append(grad_log, np.linalg.norm(xk - x_log[-1:])) fval = old_fval grad_val = grad_log[-1] return xk, fval, grad_val, x_log, y_log, grad_log
伪代码如下:
function levenberg_marquardt(fun, grad, jacobian, x0, iterations, tol):
fval = None
grad_val = None
x_log = []
y_log = []
grad_log = []
x0 = asarray(x0).flatten()
if x0.ndim == 0:
x0.shape = (1,)
k = 1
xk = x0
updateJ = 1
lamda = 0.01
old_fval = fun(x0)
gfk = grad(x0)
gnorm = np.amax(np.abs(gfk))
J = None
H = None
while (gnorm > tol) and (k < iterations):
if updateJ == 1:
x_log = np.append(x_log, xk.T)
yk = fun(xk)
y_log = np.append(y_log, yk)
J = jacobian(x0)
H = np.dot(J.T, J)
H_lm = H + (lamda * np.eye(9))
gfk = grad(xk)
pk = - np.linalg.inv(H_lm).dot(gfk)
pk = pk.A.reshape(1, -1)[0]
xk1 = xk + pk
fval = fun(xk1)
if fval < old_fval:
lamda = lamda / 10
xk = xk1
old_fval = fval
updateJ = 1
else:
updateJ = 0
lamda = lamda * 10
gnorm = np.amax(np.abs(gfk))
k = k + 1
grad_log = np.append(grad_log, np.linalg.norm(xk - x_log[-1:]))
fval = old_fval
grad_val = grad_log[-1]
return xk, fval, grad_val, x_log, y_log, grad_log