def train(self,X):
self.present_input(X)
wy,wx = self.winner_coords()
self.debug("Winner coords = "+`(wy,wx)`)
int_radius = floor(self.radius())
self.debug("Training radius = %.2f" % self.radius())
ymin = max(0,wy-int_radius)
ymax = min(wy+int_radius,self.ydim-1)
xmin = max(0,wx-int_radius)
xmax = min(wx+int_radius,self.xdim-1)
self.debug('y range = '+`(ymin,ymax)`)
self.debug('x range = '+`(xmin,xmax)`)
for y in range(ymin,ymax+1):
for x in range(xmin,xmax+1):
lattice_dist = sqrt((wx-x)**2 + (wy-y)**2)
self.debug("Trying cell %d,%d"%(x,y))
if lattice_dist <= self.radius():
self.debug("Training cell %d,%d"%(x,y))
rate = self.alpha() * gaussian(lattice_dist,self.radius())
self.weights[y,x] += rate * (X - self.weights[y,x])
self.count += 1
def gaussian_activation(self):
x = self.dists
radii = zeros(self.dists.shape) * 0.0
for u,conn_dict in enumerate(self.connections):
neighbors = take(self.weights,conn_dict.keys())
radii[u] = average(matrixnorm(neighbors-self.weights[u]))
self.__activation = gaussian(x,radii/2)
def gaussian_activation(self):
x = self.dists
radii = zeros(self.dists.shape) * 0.0
for u, conn_dict in enumerate(self.connections):
neighbors = take(self.weights, conn_dict.keys())
radii[u] = average(matrixnorm(neighbors - self.weights[u]))
self.__activation = gaussian(x, radii / 2)
def uniform_gaussian_activation(self):
x = self.dists
total = 0.0
count = 0
for u,conn_dict in enumerate(self.connections):
neighbors = take(self.weights,conn_dict.keys())
total += sum(matrixnorm(neighbors-self.weights[u]))
count += len(conn_dict)
self.__activation = gaussian(x,(total/count)/2)
def uniform_gaussian_activation(self):
x = self.dists
total = 0.0
count = 0
for u, conn_dict in enumerate(self.connections):
neighbors = take(self.weights, conn_dict.keys())
total += sum(matrixnorm(neighbors - self.weights[u]))
count += len(conn_dict)
self.__activation = gaussian(x, (total / count) / 2)
