def calculate_MI_bounded_discrete(X, Y, M_c, X_L, X_D):
get_view_index = lambda which_column: X_L['column_partition']['assignments'][which_column]
view_X = get_view_index(X)
view_Y = get_view_index(Y)
# independent
if view_X != view_Y:
return 0.0
# get cluster logps
view_state = X_L['view_state'][view_X]
cluster_logps = su.determine_cluster_crp_logps(view_state)
cluster_crps = numpy.exp(cluster_logps) # get exp'ed values for multinomial
n_clusters = len(cluster_crps)
# get X values
x_values = M_c['column_metadata'][X]['code_to_value'].values()
# get Y values
y_values = M_c['column_metadata'][Y]['code_to_value'].values()
# get components models for each cluster for columns X and Y
component_models_X = [0]*n_clusters
component_models_Y = [0]*n_clusters
for i in range(n_clusters):
cluster_models = su.create_cluster_model_from_X_L(M_c, X_L, view_X, i)
component_models_X[i] = cluster_models[X]
component_models_Y[i] = cluster_models[Y]
MI = 0.0
for x in x_values:
for y in y_values:
# calculate marginal logs
Pxy = numpy.zeros(n_clusters) # P(x,y), Joint distribution
Px = numpy.zeros(n_clusters) # P(x)
Py = numpy.zeros(n_clusters) # P(y)
# get logp of x and y in each cluster. add cluster logp's
for j in range(n_clusters):
Px[j] = component_models_X[j].calc_element_predictive_logp(x)
Py[j] = component_models_Y[j].calc_element_predictive_logp(y)
Pxy[j] = Px[j] + Py[j] + cluster_logps[j] # \sum_c P(x|c)P(y|c)P(c), Joint distribution
Px[j] += cluster_logps[j] # \sum_c P(x|c)P(c)
Py[j] += cluster_logps[j] # \sum_c P(y|c)P(c)
# sum over clusters
Px = logsumexp(Px)
Py = logsumexp(Py)
Pxy = logsumexp(Pxy)
MI += numpy.exp(Pxy)*(Pxy - (Px + Py))
# ignore MI < 0
if MI <= 0.0:
MI = 0.0
return MI
def calculate_MI_bounded_discrete(X, Y, M_c, X_L, X_D):
get_view_index = lambda which_column: X_L['column_partition']['assignments'][which_column]
view_X = get_view_index(X)
view_Y = get_view_index(Y)
# independent
if view_X != view_Y:
return 0.0
# get cluster logps
view_state = X_L['view_state'][view_X]
cluster_logps = su.determine_cluster_crp_logps(view_state)
cluster_crps = numpy.exp(cluster_logps) # get exp'ed values for multinomial
n_clusters = len(cluster_crps)
# get X values
x_values = M_c['column_metadata'][X]['code_to_value'].values()
# get Y values
y_values = M_c['column_metadata'][Y]['code_to_value'].values()
# get components models for each cluster for columns X and Y
component_models_X = [0]*n_clusters
component_models_Y = [0]*n_clusters
for i in range(n_clusters):
cluster_models = su.create_cluster_model_from_X_L(M_c, X_L, view_X, i)
component_models_X[i] = cluster_models[X]
component_models_Y[i] = cluster_models[Y]
MI = 0.0
for x in x_values:
for y in y_values:
# calculate marginal logs
Pxy = numpy.zeros(n_clusters) # P(x,y), Joint distribution
Px = numpy.zeros(n_clusters) # P(x)
Py = numpy.zeros(n_clusters) # P(y)
# get logp of x and y in each cluster. add cluster logp's
for j in range(n_clusters):
Px[j] = component_models_X[j].calc_element_predictive_logp(x)
Py[j] = component_models_Y[j].calc_element_predictive_logp(y)
Pxy[j] = Px[j] + Py[j] + cluster_logps[j] # \sum_c P(x|c)P(y|c)P(c), Joint distribution
Px[j] += cluster_logps[j] # \sum_c P(x|c)P(c)
Py[j] += cluster_logps[j] # \sum_c P(y|c)P(c)
# sum over clusters
Px = logsumexp(Px)
Py = logsumexp(Py)
Pxy = logsumexp(Pxy)
MI += numpy.exp(Pxy)*(Pxy - (Px + Py))
# ignore MI < 0
if MI <= 0.0:
MI = 0.0
return MI
def sample_from_view(M_c, X_L, X_D, get_next_seed):
view_col = X_L['column_partition']['assignments'][0]
view_col2 = X_L['column_partition']['assignments'][1]
same_view = True
if view_col2 != view_col:
same_view = False
view_state = X_L['view_state'][view_col]
view_state2 = X_L['view_state'][view_col2]
cluster_crps = numpy.exp(su.determine_cluster_crp_logps(view_state))
cluster_crps2 = numpy.exp(su.determine_cluster_crp_logps(view_state2))
assert (math.fabs(numpy.sum(cluster_crps) - 1) < .00000001)
samples = numpy.zeros((n, 2))
cluster_idx1 = numpy.nonzero(numpy.random.multinomial(1,
cluster_crps))[0][0]
cluster_model1 = su.create_cluster_model_from_X_L(M_c, X_L, view_col,
cluster_idx1)
if same_view:
cluster_idx2 = cluster_idx1
cluster_model2 = cluster_model1
else:
cluster_idx2 = numpy.nonzero(numpy.random.multinomial(
1, cluster_crps2))[0][0]
cluster_model2 = su.create_cluster_model_from_X_L(
M_c, X_L, view_col2, cluster_idx2)
component_model1 = cluster_model1[0]
x = component_model1.get_draw(get_next_seed())
component_model2 = cluster_model2[1]
y = component_model2.get_draw(get_next_seed())
return x, y
def sample_from_view(M_c, X_L, X_D, get_next_seed):
view_col = X_L['column_partition']['assignments'][0]
view_col2 = X_L['column_partition']['assignments'][1]
same_view = True
if view_col2 != view_col:
same_view = False
view_state = X_L['view_state'][view_col]
view_state2 = X_L['view_state'][view_col2]
cluster_crps = numpy.exp(su.determine_cluster_crp_logps(view_state))
cluster_crps2 = numpy.exp(su.determine_cluster_crp_logps(view_state2))
assert( math.fabs(numpy.sum(cluster_crps) - 1) < .00000001 )
samples = numpy.zeros((n,2))
cluster_idx1 = numpy.nonzero(numpy.random.multinomial(1, cluster_crps))[0][0]
cluster_model1 = su.create_cluster_model_from_X_L(M_c, X_L, view_col, cluster_idx1)
if same_view:
cluster_idx2 = cluster_idx1
cluster_model2 = cluster_model1
else:
cluster_idx2 = numpy.nonzero(numpy.random.multinomial(1, cluster_crps2))[0][0]
cluster_model2 = su.create_cluster_model_from_X_L(M_c, X_L, view_col2, cluster_idx2)
component_model1 = cluster_model1[0]
x = component_model1.get_draw(get_next_seed())
component_model2 = cluster_model2[1]
y = component_model2.get_draw(get_next_seed())
return x, y
def estimiate_MI_sample(X, Y, M_c, X_L, X_D, get_next_seed, n_samples=1000):
get_view_index = lambda which_column: X_L['column_partition'][
'assignments'][which_column]
view_X = get_view_index(X)
view_Y = get_view_index(Y)
# independent
if view_X != view_Y:
return 0.0
# get cluster logps
view_state = X_L['view_state'][view_X]
cluster_logps = su.determine_cluster_crp_logps(view_state)
cluster_crps = numpy.exp(
cluster_logps) # get exp'ed values for multinomial
n_clusters = len(cluster_crps)
# get components models for each cluster for columns X and Y
component_models_X = [0] * n_clusters
component_models_Y = [0] * n_clusters
for i in range(n_clusters):
cluster_models = su.create_cluster_model_from_X_L(M_c, X_L, view_X, i)
component_models_X[i] = cluster_models[X]
component_models_Y[i] = cluster_models[Y]
MI = 0.0 # mutual information
for _ in range(n_samples):
# draw a cluster
cluster_idx = numpy.nonzero(numpy.random.multinomial(
1, cluster_crps))[0][0]
# get a sample from each cluster
x = component_models_X[cluster_idx].get_draw(get_next_seed())
y = component_models_Y[cluster_idx].get_draw(get_next_seed())
# calculate marginal logs
Pxy = numpy.zeros(n_clusters) # P(x,y), Joint distribution
Px = numpy.zeros(n_clusters) # P(x)
Py = numpy.zeros(n_clusters) # P(y)
# get logp of x and y in each cluster. add cluster logp's
for j in range(n_clusters):
Px[j] = component_models_X[j].calc_element_predictive_logp(x)
Py[j] = component_models_Y[j].calc_element_predictive_logp(y)
Pxy[j] = Px[j] + Py[j] + cluster_logps[
j] # \sum_c P(x|c)P(y|c)P(c), Joint distribution
Px[j] += cluster_logps[j] # \sum_c P(x|c)P(c)
Py[j] += cluster_logps[j] # \sum_c P(y|c)P(c)
# pdb.set_trace()
# sum over clusters
Px = logsumexp(Px)
Py = logsumexp(Py)
Pxy = logsumexp(Pxy)
# add to MI
MI += Pxy - (Px + Py)
# average
MI /= float(n_samples)
# ignore MI < 0
if MI <= 0.0:
MI = 0.0
return MI
def estimiate_MI_sample_hist(X,
Y,
M_c,
X_L,
X_D,
get_next_seed,
n_samples=10000):
get_view_index = lambda which_column: X_L['column_partition'][
'assignments'][which_column]
view_X = get_view_index(X)
view_Y = get_view_index(Y)
# independent
if view_X != view_Y:
return 0.0
# get cluster logps
view_state = X_L['view_state'][view_X]
cluster_logps = su.determine_cluster_crp_logps(view_state)
cluster_crps = numpy.exp(cluster_logps)
n_clusters = len(cluster_crps)
# get components models for each cluster for columns X and Y
component_models_X = [0] * n_clusters
component_models_Y = [0] * n_clusters
for i in range(n_clusters):
cluster_models = su.create_cluster_model_from_X_L(M_c, X_L, view_X, i)
component_models_X[i] = cluster_models[X]
component_models_Y[i] = cluster_models[Y]
MI = 0.0
samples = numpy.zeros((n_samples, 2), dtype=float)
samples_x = numpy.zeros(n_samples, dtype=float)
samples_y = numpy.zeros(n_samples, dtype=float)
# draw the samples
for i in range(n_samples):
# draw a cluster
cluster_idx = numpy.nonzero(numpy.random.multinomial(
1, cluster_crps))[0][0]
x = component_models_X[cluster_idx].get_draw(get_next_seed())
y = component_models_Y[cluster_idx].get_draw(get_next_seed())
samples[i, 0] = x
samples[i, 1] = y
samples_x[i] = x
samples_y[i] = y
# calculate the number of bins and ranges
N = float(n_samples)
r, _ = corr(samples_x, samples_y)
k = round(.5 + .5 * (1 + 4 * ((6 * N * r**2.) / (1 - r**2.))**.5)**.5) + 1
sigma_x = numpy.std(samples_x)
mu_x = numpy.mean(samples_x)
sigma_y = numpy.std(samples_y)
mu_y = numpy.mean(samples_y)
range_x = numpy.linspace(mu_x - 3. * sigma_x, mu_x + 3 * sigma_x, k)
range_y = numpy.linspace(mu_y - 3. * sigma_y, mu_y + 3 * sigma_y, k)
PXY, _, _ = numpy.histogram2d(samples[:, 0],
samples[:, 1],
bins=[range_x, range_y])
PX, _ = numpy.histogram(samples_x, bins=range_x)
PY, _ = numpy.histogram(samples_y, bins=range_y)
MI = 0
for i_x in range(PXY.shape[0]):
for i_y in range(PXY.shape[1]):
Pxy = PXY[i_x, i_y]
Px = PX[i_x]
Py = PY[i_y]
if Pxy > 0.0 and Px > 0.0 and Py > 0.0:
MI += (Pxy / N) * math.log(Pxy * N / (Px * Py))
# ignore MI < 0
if MI <= 0.0:
MI = 0.0
return MI
def calculate_MI_bounded_discrete(X, Y, M_c, X_L, _X_D):
get_view_index = lambda which_column: X_L['column_partition']['assignments'][which_column]
view_X = get_view_index(X)
view_Y = get_view_index(Y)
# independent
if view_X != view_Y:
return 0.0
# get cluster logps
view_state = X_L['view_state'][view_X]
cluster_logps = numpy.array(su.determine_cluster_crp_logps(view_state))
n_clusters = len(cluster_logps)
# get X values
x_values = M_c['column_metadata'][X]['code_to_value'].values()
# get Y values
y_values = M_c['column_metadata'][Y]['code_to_value'].values()
# get components models for each cluster for columns X and Y
component_models_X = [0]*n_clusters
component_models_Y = [0]*n_clusters
for i in range(n_clusters):
cluster_models = su.create_cluster_model_from_X_L(M_c, X_L, view_X, i)
component_models_X[i] = cluster_models[X]
component_models_Y[i] = cluster_models[Y]
def marginal_predictive_logps_by_cluster(value, component_models):
return numpy.array([
component_models[j].calc_element_predictive_logp(value)
+ cluster_logps[j]
for j in range(n_clusters)])
x_marginal_predictive_logps_by_cluster = \
[marginal_predictive_logps_by_cluster(x, component_models_X)
for x in x_values]
# \sum_c P(x|c)P(c)
x_net_marginal_predictive_logps = \
[logsumexp(ps) for ps in x_marginal_predictive_logps_by_cluster]
y_marginal_predictive_logps_by_cluster = \
[marginal_predictive_logps_by_cluster(y, component_models_Y)
for y in y_values]
# \sum_c P(y|c)P(c)
y_net_marginal_predictive_logps = \
[logsumexp(ps) for ps in y_marginal_predictive_logps_by_cluster]
MI = 0.0
for (i,x) in enumerate(x_values):
x_marginals = x_marginal_predictive_logps_by_cluster[i]
for (j,y) in enumerate(y_values):
y_marginals = y_marginal_predictive_logps_by_cluster[j]
# cluster prob is double-counted in sum of marginals
joint_predictive_logp_by_cluster = \
x_marginals + y_marginals - cluster_logps
# \sum_c P(x|c)P(y|c)P(c), Joint distribution
joint_predictive_logp = logsumexp(joint_predictive_logp_by_cluster)
MI += math.exp(joint_predictive_logp) * \
(joint_predictive_logp - \
(x_net_marginal_predictive_logps[i] + \
y_net_marginal_predictive_logps[j]))
# ignore MI < 0
if MI <= 0.0:
MI = 0.0
return MI
def estimate_MI_sample(X, Y, M_c, X_L, _X_D, get_next_seed, n_samples=1000):
random_state = numpy.random.RandomState(get_next_seed())
get_view_index = lambda which_column: X_L['column_partition']['assignments'][which_column]
view_X = get_view_index(X)
view_Y = get_view_index(Y)
# independent
if view_X != view_Y:
return 0.0
# get cluster logps
view_state = X_L['view_state'][view_X]
cluster_logps = su.determine_cluster_crp_logps(view_state)
cluster_crps = numpy.exp(cluster_logps) # get exp'ed values for multinomial
n_clusters = len(cluster_crps)
# get components models for each cluster for columns X and Y
component_models_X = [0]*n_clusters
component_models_Y = [0]*n_clusters
for i in range(n_clusters):
cluster_models = su.create_cluster_model_from_X_L(M_c, X_L, view_X, i)
component_models_X[i] = cluster_models[X]
component_models_Y[i] = cluster_models[Y]
# MI = 0.0 # mutual information
MI = numpy.zeros(n_samples)
weights = numpy.zeros(n_samples)
for i in range(n_samples):
# draw a cluster
cluster_idx = numpy.nonzero(random_state.multinomial(1, cluster_crps))[0][0]
# get a sample from each cluster
x = component_models_X[cluster_idx].get_draw(get_next_seed())
y = component_models_Y[cluster_idx].get_draw(get_next_seed())
# calculate marginal logs
Pxy = numpy.zeros(n_clusters) # P(x,y), Joint distribution
Px = numpy.zeros(n_clusters) # P(x)
Py = numpy.zeros(n_clusters) # P(y)
# get logp of x and y in each cluster. add cluster logp's
for j in range(n_clusters):
Px[j] = component_models_X[j].calc_element_predictive_logp(x)
Py[j] = component_models_Y[j].calc_element_predictive_logp(y)
Pxy[j] = Px[j] + Py[j] + cluster_logps[j] # \sum_c P(x|c)P(y|c)P(c), Joint distribution
Px[j] += cluster_logps[j] # \sum_c P(x|c)P(c)
Py[j] += cluster_logps[j] # \sum_c P(y|c)P(c)
# pdb.set_trace()
# sum over clusters
Px = logsumexp(Px)
Py = logsumexp(Py)
Pxy = logsumexp(Pxy)
# add to MI
# MI += Pxy - (Px + Py)
MI[i] = Pxy - (Px + Py)
weights[i] = Pxy
# do weighted average with underflow protection
# MI /= float(n_samples)
Z = logsumexp(weights)
weights = numpy.exp(weights-Z)
MI_ret = numpy.sum(MI*weights)
# ignore MI < 0
if MI_ret <= 0.0:
MI_ret = 0.0
return MI_ret
def estimiate_MI_sample_hist(X, Y, M_c, X_L, X_D, get_next_seed, n_samples=10000):
get_view_index = lambda which_column: X_L['column_partition']['assignments'][which_column]
view_X = get_view_index(X)
view_Y = get_view_index(Y)
# independent
if view_X != view_Y:
return 0.0
# get cluster logps
view_state = X_L['view_state'][view_X]
cluster_logps = su.determine_cluster_crp_logps(view_state)
cluster_crps = numpy.exp(cluster_logps)
n_clusters = len(cluster_crps)
# get components models for each cluster for columns X and Y
component_models_X = [0]*n_clusters
component_models_Y = [0]*n_clusters
for i in range(n_clusters):
cluster_models = su.create_cluster_model_from_X_L(M_c, X_L, view_X, i)
component_models_X[i] = cluster_models[X]
component_models_Y[i] = cluster_models[Y]
MI = 0.0
samples = numpy.zeros((n_samples,2), dtype=float)
samples_x = numpy.zeros(n_samples, dtype=float)
samples_y = numpy.zeros(n_samples, dtype=float)
# draw the samples
for i in range(n_samples):
# draw a cluster
cluster_idx = numpy.nonzero(numpy.random.multinomial(1, cluster_crps))[0][0]
x = component_models_X[cluster_idx].get_draw(get_next_seed())
y = component_models_Y[cluster_idx].get_draw(get_next_seed())
samples[i,0] = x
samples[i,1] = y
samples_x[i] = x
samples_y[i] = y
# calculate the number of bins and ranges
N = float(n_samples)
r,_ = corr(samples_x, samples_y)
k = round(.5+.5*(1+4*((6*N*r**2.)/(1-r**2.))**.5)**.5)+1
sigma_x = numpy.std(samples_x)
mu_x = numpy.mean(samples_x)
sigma_y = numpy.std(samples_y)
mu_y = numpy.mean(samples_y)
range_x = numpy.linspace(mu_x-3.*sigma_x,mu_x+3*sigma_x,k)
range_y = numpy.linspace(mu_y-3.*sigma_y,mu_y+3*sigma_y,k)
PXY, _, _ = numpy.histogram2d(samples[:,0], samples[:,1], bins=[range_x,range_y])
PX,_ = numpy.histogram(samples_x,bins=range_x)
PY,_ = numpy.histogram(samples_y,bins=range_y)
MI = 0
for i_x in range(PXY.shape[0]):
for i_y in range(PXY.shape[1]):
Pxy = PXY[i_x,i_y]
Px = PX[i_x]
Py = PY[i_y]
if Pxy > 0.0 and Px > 0.0 and Py > 0.0:
MI += (Pxy/N)*math.log(Pxy*N/(Px*Py))
# ignore MI < 0
if MI <= 0.0:
MI = 0.0
return MI
def calculate_MI_bounded_discrete(X, Y, M_c, X_L, _X_D):
get_view_index = lambda which_column: X_L['column_partition']['assignments'][which_column]
view_X = get_view_index(X)
view_Y = get_view_index(Y)
# independent
if view_X != view_Y:
return 0.0
# get cluster logps
view_state = X_L['view_state'][view_X]
cluster_logps = numpy.array(su.determine_cluster_crp_logps(view_state))
n_clusters = len(cluster_logps)
# get X values
x_values = M_c['column_metadata'][X]['code_to_value'].values()
# get Y values
y_values = M_c['column_metadata'][Y]['code_to_value'].values()
# get components models for each cluster for columns X and Y
component_models_X = [0]*n_clusters
component_models_Y = [0]*n_clusters
for i in range(n_clusters):
cluster_models = su.create_cluster_model_from_X_L(M_c, X_L, view_X, i)
component_models_X[i] = cluster_models[X]
component_models_Y[i] = cluster_models[Y]
def marginal_predictive_logps_by_cluster(value, component_models):
return numpy.array([
component_models[j].calc_element_predictive_logp(value)
+ cluster_logps[j]
for j in range(n_clusters)])
x_marginal_predictive_logps_by_cluster = \
[marginal_predictive_logps_by_cluster(x, component_models_X)
for x in x_values]
# \sum_c P(x|c)P(c)
x_net_marginal_predictive_logps = \
[logsumexp(ps) for ps in x_marginal_predictive_logps_by_cluster]
y_marginal_predictive_logps_by_cluster = \
[marginal_predictive_logps_by_cluster(y, component_models_Y)
for y in y_values]
# \sum_c P(y|c)P(c)
y_net_marginal_predictive_logps = \
[logsumexp(ps) for ps in y_marginal_predictive_logps_by_cluster]
MI = 0.0
for (i,x) in enumerate(x_values):
x_marginals = x_marginal_predictive_logps_by_cluster[i]
for (j,y) in enumerate(y_values):
y_marginals = y_marginal_predictive_logps_by_cluster[j]
# cluster prob is double-counted in sum of marginals
joint_predictive_logp_by_cluster = \
x_marginals + y_marginals - cluster_logps
# \sum_c P(x|c)P(y|c)P(c), Joint distribution
joint_predictive_logp = logsumexp(joint_predictive_logp_by_cluster)
MI += math.exp(joint_predictive_logp) * \
(joint_predictive_logp - \
(x_net_marginal_predictive_logps[i] + \
y_net_marginal_predictive_logps[j]))
# ignore MI < 0
if MI <= 0.0:
MI = 0.0
return MI