Python numpy.matrix方法代碼示例

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def set_state(self, new_state):
        '''
        Set Hipster's new type.
        This comes from evaling the env.
        Args:
            new_state: the agent's new state
        Returns:
            None
        '''
        old_type = self.ntype
        self.state = new_state
        self.ntype = STATE_MAP[new_state]
        self.env.change_agent_type(self, old_type, self.ntype)
        # Unlike the forest fire model, we don't here update environment's cell's
        # transition matrices. This is because the cell's transition matrix depends
        # on all the agents near it, not just one. 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def get_pre(self, agent, n_census):

        trans_str = ""

        d, total = self.dir_info(agent)

        if type(agent) == Zombie:
            trans_str += self.zombie_trans(d, total)
        else:
            trans_str += self.human_trans(d, total)

        trans_matrix = markov.from_matrix(np.matrix(trans_str))
        return trans_matrix
       
    # Finds out which direction (NORTH, SOUTH, EAST, WEST) as more of the 
    # opposite agent type depending on what agent we are dealing with
    # CAN'T GET RID OF OLD METHOD OF MOVEMENT DUE TO ERRORS
    # IN OTHER SCRIPTS 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def predict_all(X, all_theta):
    rows = X.shape[0]
    params = X.shape[1]
    num_labels = all_theta.shape[0]
    
    # same as before, insert ones to match the shape
    X = np.insert(X, 0, values=np.ones(rows), axis=1)
    
    # convert to matrices
    X = np.matrix(X)
    all_theta = np.matrix(all_theta)
    
    # compute the class probability for each class on each training instance
    h = sigmoid(X * all_theta.T)
    
    # create array of the index with the maximum probability
    h_argmax = np.argmax(h, axis=1)
    
    # because our array was zero-indexed we need to add one for the true label prediction
    h_argmax = h_argmax + 1
    
    return h_argmax 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def cost0(params, input_size, hidden_size, num_labels, X, y, learning_rate):
    m = X.shape[0]
    X = np.matrix(X)
    y = np.matrix(y)
    
    # reshape the parameter array into parameter matrices for each layer
    theta1 = np.matrix(np.reshape(params[:hidden_size * (input_size + 1)], (hidden_size, (input_size + 1))))
    theta2 = np.matrix(np.reshape(params[hidden_size * (input_size + 1):], (num_labels, (hidden_size + 1))))
    
    # run the feed-forward pass
    a1, z2, a2, z3, h = forward_propagate(X, theta1, theta2)
    
    # compute the cost
    J = 0
    for i in range(m):
        first_term = np.multiply(-y[i,:], np.log(h[i,:]))
        second_term = np.multiply((1 - y[i,:]), np.log(1 - h[i,:]))
        J += np.sum(first_term - second_term)
    
    J = J / m
    
    return J 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def cost(params, Y, R, num_features):
    Y = np.matrix(Y)  # (1682, 943)
    R = np.matrix(R)  # (1682, 943)
    num_movies = Y.shape[0]
    num_users = Y.shape[1]
    
    # reshape the parameter array into parameter matrices
    X = np.matrix(np.reshape(params[:num_movies * num_features], (num_movies, num_features)))  # (1682, 10)
    Theta = np.matrix(np.reshape(params[num_movies * num_features:], (num_users, num_features)))  # (943, 10)
    
    # initializations
    J = 0
    
    # compute the cost
    error = np.multiply((X * Theta.T) - Y, R)  # (1682, 943)
    squared_error = np.power(error, 2)  # (1682, 943)
    J = (1. / 2) * np.sum(squared_error)
    
    return J 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def gradientReg(theta, X, y, learningRate):
    theta = np.matrix(theta)
    X = np.matrix(X)
    y = np.matrix(y)
    
    parameters = int(theta.ravel().shape[1])
    grad = np.zeros(parameters)
    
    error = sigmoid(X * theta.T) - y
    
    for i in range(parameters):
        term = np.multiply(error, X[:,i])
        
        if (i == 0):
            grad[i] = np.sum(term) / len(X)
        else:
            grad[i] = (np.sum(term) / len(X)) + ((learningRate / len(X)) * theta[:,i])
    
    return grad 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def _icp_find_rigid_transform(p_from, p_target):
	######### flag = 1 for SVD 
	######### flag = 2 for Gaussian
	A, B = np.copy(p_from), np.copy(p_target)

	centroid_A = np.mean(A, axis=0)
	centroid_B = np.mean(B, axis=0)

	A -= centroid_A						# 500x3
	B -= centroid_B						# 500x3

	H = np.dot(A.T, B)					# 3x3
	# print(H)
	U, S, Vt = np.linalg.svd(H)
	# H_new = U*np.matrix(np.diag(S))*Vt
	# print(H_new)
	R = np.dot(Vt.T, U.T)

	# special reflection case
	if np.linalg.det(R) < 0:
		Vt[2,:] *= -1
		R = np.dot(Vt.T, U.T)
	t = np.dot(-R, centroid_A) + centroid_B
	return R, t 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def controller_lqr_discrete_from_continuous_time(A, B, Q, R, dt):
    """Solve the discrete time LQR controller for a continuous time system.
    
    A and B are system matrices, describing the systems dynamics:
     dx/dt = A x + B u
    
    The controller minimizes the infinite horizon quadratic cost function:
     cost = integral (x.T*Q*x + u.T*R*u) dt
    where Q is a positive semidefinite matrix, and R is positive definite matrix.
    
    The controller is implemented to run at discrete times, at a rate given by
     onboard_dt, 
    i.e. u[k] = -K*x(k*t)
    Discretization is done by zero order hold.
    
    Returns K, X, eigVals:
    Returns gain the optimal gain K, the solution matrix X, and the closed loop system eigenvalues.
    The optimal input is then computed as:
     input: u = -K*x
    """
    #ref Bertsekas, p.151
    
    Ad, Bd = analysis.discretise_time(A, B, dt)

    return controller_lqr_discrete_time(Ad, Bd, Q, R) 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def heston_construct_correlated_path(params: ModelParameters,
                                     brownian_motion_one: np.array):
    """
    This method is a simplified version of the Cholesky decomposition method
    for just two assets. It does not make use of matrix algebra and is therefore
    quite easy to implement.

    Arguments:
        params : ModelParameters
            The parameters for the stochastic model.
        brownian_motion_one : np.array
            (Not filled)

    Returns:
        A correlated brownian motion path.
    """
    # We do not multiply by sigma here, we do that in the Heston model
    sqrt_delta = np.sqrt(params.all_delta)
    # Construct a path correlated to the first path
    brownian_motion_two = []
    for i in range(params.all_time - 1):
        term_one = params.cir_rho * brownian_motion_one[i]
        term_two = np.sqrt(1 - pow(params.cir_rho, 2)) * random.normalvariate(0, sqrt_delta)
        brownian_motion_two.append(term_one + term_two)
    return np.array(brownian_motion_one), np.array(brownian_motion_two) 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def add_pixels(self, uv_px, img1d, weight=None):
        # Lookup row & column for each in-bounds coordinate.
        mask = self.get_mask(uv_px)
        xx = uv_px[0,mask]
        yy = uv_px[1,mask]
        # Update matrix according to assigned weight.
        if weight is None:
            img1d[mask] = self.img[yy,xx]
        elif np.isscalar(weight):
            img1d[mask] += self.img[yy,xx] * weight
        else:
            w1 = np.asmatrix(weight, dtype='float32')
            w3 = w1.transpose() * np.ones((1,3))
            img1d[mask] += np.multiply(self.img[yy,xx], w3[mask])


# A panorama image made from several FisheyeImage sources.
# TODO: Add support for supersampled anti-aliasing filters. 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def render_cubemap(self, out_size, mode='blend'):
        # Create coordinate arrays.
        cvec = np.arange(out_size, dtype='float32') - out_size/2        # Coordinate range [-S/2, S/2)
        vec0 = np.ones(out_size*out_size, dtype='float32') * out_size/2 # Constant vector +S/2
        vec1 = np.repeat(cvec, out_size)                                # Increment every N steps
        vec2 = np.tile(cvec, out_size)                                  # Sweep N times
        # Create XYZ coordinate vectors and render each cubemap face.
        render = lambda(xyz): self._render(xyz, out_size, out_size, mode)
        xm = render(np.matrix([-vec0, vec1, vec2]))     # -X face
        xp = render(np.matrix([vec0, vec1, -vec2]))     # +X face
        ym = render(np.matrix([-vec1, -vec0, vec2]))    # -Y face
        yp = render(np.matrix([vec1, vec0, vec2]))      # +Y face
        zm = render(np.matrix([-vec2, vec1, -vec0]))    # -Z face
        zp = render(np.matrix([vec2, vec1, vec0]))      # +Z face
        # Concatenate the individual faces in canonical order:
        # https://en.wikipedia.org/wiki/Cube_mapping#Memory_Addressing
        img_mat = np.concatenate([zp, zm, ym, yp, xm, xp], axis=0)
        return Image.fromarray(img_mat)

    # Get XYZ vectors for an equirectangular render, in raster order.
    # (Each row left to right, with rows concatenates from top to bottom.) 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def _get_equirectangular_raster(self, out_size):
        # Set image size (2x1 aspect ratio)
        rows = out_size
        cols = 2*out_size
        # Calculate longitude of each column.
        theta_x = np.linspace(-pi, pi, cols, endpoint=False, dtype='float32')
        cos_x = np.cos(theta_x).reshape(1,cols)
        sin_x = np.sin(theta_x).reshape(1,cols)
        # Calculate lattitude of each row.
        ystep = pi / rows
        theta_y = np.linspace(-pi/2 + ystep/2, pi/2 - ystep/2, rows, dtype='float32')
        cos_y = np.cos(theta_y).reshape(rows,1)
        sin_y = np.sin(theta_y).reshape(rows,1)
        # Calculate X, Y, and Z coordinates for each output pixel.
        x = cos_y * cos_x
        y = sin_y * np.ones((1,cols), dtype='float32')
        z = cos_y * sin_x
        # Vectorize the coordinates in raster order.
        xyz = np.matrix([x.ravel(), y.ravel(), z.ravel()])
        return [xyz, rows, cols]

    # Convert all lens parameters to a state vector. See also: optimize() 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def create_iden_matrix(n):
    """
    Create a dim1 * dim2 identity matrix.

    Returns: the new matrix.
    """
    matrix_init = [[] for i in range(n)]
    for i in range(0, n): 
        for j in range(0, n): 
            if i == j:
                matrix_init[i].append(1)
            else:
                matrix_init[i].append(0)

    return np.matrix(matrix_init) 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def state_vector(vlen, init_state):
    vals = ""
    for i in range(vlen):
        if i == init_state:
            vals = vals + "1 "
        else:
            vals = vals + "0 "
    return np.matrix(vals) 

# 需要導入模塊: import numpy [as 別名]
# 或者: from numpy import matrix [as 別名]
def from_matrix(m):
    """
    Takes an numpy matrix and returns a prehension.
    """
    pre = MarkovPre("")
    pre.matrix = m
    return pre