def GPDM_solver(self):
train_data = []
print(self.gene_data)
for key_gene in self.select_gene:
train_data.append(self.gene_data[key_gene])
self.train_data = np.array(train_data).T
output = GPLVM(self.train_data, self.latent_dim, init='PCA')
output.optimize(messages=True, max_iters=20)
return output
def plot_latent(self, *args, **kwargs):
input_1, input_2 = GPLVM.plot_latent(*args, **kwargs)
pb.plot(m.Z[:, input_1], m.Z[:, input_2], '^w')
def plot(self):
GPLVM.plot(self)
# passing Z without a small amout of jitter will induce the white kernel where we don;t want it!
mu, var, upper, lower = SparseGPRegression.predict(self, self.Z + np.random.randn(*self.Z.shape) * 0.0001)
pb.plot(mu[:, 0] , mu[:, 1], 'ko')
def plot_latent(self, *args, **kwargs):
input_1, input_2 = GPLVM.plot_latent(*args, **kwargs)
pb.plot(m.Z[:, input_1], m.Z[:, input_2], '^w')
def plot(self):
GPLVM.plot(self)
# passing Z without a small amout of jitter will induce the white kernel where we don;t want it!
mu, var, upper, lower = SparseGPRegression.predict(self, self.Z + np.random.randn(*self.Z.shape) * 0.0001)
pb.plot(mu[:, 0] , mu[:, 1], 'ko')
plt.tick_params(bottom=False,
left=False,
right=False,
top=False)
plt.gca().spines['right'].set_visible(False)
plt.gca().spines['top'].set_visible(False)
plt.gca().spines['left'].set_visible(False)
plt.gca().spines['bottom'].set_visible(False)
plt.savefig(f_name)
if __name__ == "__main__":
np.random.seed(3)
image = np.asarray(Image.open('./data/shobon.png').convert('L'))
# binarization
a = np.where(image < 240)
shobon = np.c_[a[1], np.flipud(a[0])]
img_save(shobon, f_name='shobon.png')
# high dimension
dim = 100
w = np.random.normal(size=2*dim).reshape(2, dim)
high = shobon @ w
high = high + np.random.normal(size=high.shape[0]*high.shape[1]).reshape(high.shape[0], high.shape[1])
# latent
gplvm = GPLVM(high, 2)
latent = gplvm.X
img_save(latent, f_name='latent.png')
