def sample_tree(data, settings, param, cache, cache_tmp):
p = TreeMCMC(range(data['n_train']), param, settings, cache_tmp)
grow_nodes = [0]
while grow_nodes:
node_id = grow_nodes.pop(0)
p.depth = max(p.depth, get_depth(node_id))
log_psplit = np.log(p.compute_psplit(node_id, param))
train_ids = p.train_ids[node_id]
(do_not_split_node_id, feat_id_chosen, split_chosen, idx_split_global, log_sis_ratio, logprior_nodeid, \
train_ids_left, train_ids_right, cache_tmp, loglik_left, loglik_right) \
= p.precomputed_proposal(data, param, settings, cache, node_id, train_ids, log_psplit)
if do_not_split_node_id:
p.do_not_split[node_id] = True
else:
p.update_left_right_statistics(cache_tmp, node_id, logprior_nodeid, train_ids_left,\
train_ids_right, loglik_left, loglik_right, feat_id_chosen, split_chosen, \
idx_split_global, settings, param, data, cache)
left, right = get_children_id(node_id)
grow_nodes.append(left)
grow_nodes.append(right)
# create mcmc structures
p.both_children_terminal.append(node_id)
parent = get_parent_id(node_id)
if (node_id != 0) and (parent in p.non_leaf_nodes):
p.inner_pc_pairs.append((parent, node_id))
if node_id != 0:
try:
p.both_children_terminal.remove(parent)
except ValueError:
pass
if settings.debug:
print 'sampled new tree:'
p.print_tree()
return p
def compute_log_inv_acc_p(self, node_id, param, len_both_children_terminal, loglik, grow_nodes, \
cache, settings, data):
# 1/acc for PRUNE is acc for GROW except for corrections to both_children_terminal
# and grow_nodes list
logprior_children = 0.0
left, right = get_children_id(node_id)
if not no_valid_split_exists(data, cache, self.train_ids[left], settings):
logprior_children += np.log(self.compute_pnosplit(left, param))
if not no_valid_split_exists(data, cache, self.train_ids[right], settings):
logprior_children += np.log(self.compute_pnosplit(right, param))
try:
check_if_zero(logprior_children - self.logprior[left] - self.logprior[right])
except AssertionError:
print 'oh oh ... looks like a bug in compute_log_inv_acc_p'
print 'term 1 = %s' % logprior_children
print 'term 2 = %s, 2a = %s, 2b = %s' % (self.logprior[left]+self.logprior[right], \
self.logprior[left], self.logprior[right])
print 'node_id = %s, left = %s, right = %s, logprior = %s' % (node_id, left, right, self.logprior)
raise AssertionError
log_inv_acc_prior = np.log(self.compute_psplit(node_id, param)) \
- np.log(self.compute_pnosplit(node_id, param)) \
-np.log(len_both_children_terminal) + np.log(len(grow_nodes)) \
+ logprior_children
log_inv_acc_loglik = (loglik - self.loglik[node_id])
log_inv_acc = log_inv_acc_loglik + log_inv_acc_prior
if settings.verbose >= 2:
print 'compute_log_inv_acc_p: log_acc_loglik = %s, log_acc_prior = %s' \
% (-log_inv_acc_loglik, -log_inv_acc_prior)
assert(log_inv_acc > -np.inf)
return log_inv_acc
def compute_log_inv_acc_p(self, node_id, param, len_both_children_terminal, loglik, grow_nodes, \
cache, settings, data):
# 1/acc for PRUNE is acc for GROW except for corrections to both_children_terminal
# and grow_nodes list
logprior_children = 0.0
left, right = get_children_id(node_id)
if not no_valid_split_exists(data, cache, self.train_ids[left], settings):
logprior_children += np.log(self.compute_pnosplit(left, param))
if not no_valid_split_exists(data, cache, self.train_ids[right], settings):
logprior_children += np.log(self.compute_pnosplit(right, param))
try:
check_if_zero(logprior_children - self.logprior[left] - self.logprior[right])
except AssertionError:
print 'oh oh ... looks like a bug in compute_log_inv_acc_p'
print 'term 1 = %s' % logprior_children
print 'term 2 = %s, 2a = %s, 2b = %s' % (self.logprior[left]+self.logprior[right], \
self.logprior[left], self.logprior[right])
print 'node_id = %s, left = %s, right = %s, logprior = %s' % (node_id, left, right, self.logprior)
raise AssertionError
log_inv_acc_prior = np.log(self.compute_psplit(node_id, param)) \
- np.log(self.compute_pnosplit(node_id, param)) \
-np.log(len_both_children_terminal) + np.log(len(grow_nodes)) \
+ logprior_children
log_inv_acc_loglik = (loglik - self.loglik[node_id])
log_inv_acc = log_inv_acc_loglik + log_inv_acc_prior
if settings.verbose >= 2:
print 'compute_log_inv_acc_p: log_acc_loglik = %s, log_acc_prior = %s' \
% (-log_inv_acc_loglik, -log_inv_acc_prior)
assert(log_inv_acc > -np.inf)
return log_inv_acc
def sample_tree(data, settings, param, cache, cache_tmp):
p = TreeMCMC(range(data['n_train']), param, settings, cache_tmp)
grow_nodes = [0]
while grow_nodes:
node_id = grow_nodes.pop(0)
p.depth = max(p.depth, get_depth(node_id))
log_psplit = np.log(p.compute_psplit(node_id, param))
train_ids = p.train_ids[node_id]
(do_not_split_node_id, feat_id_chosen, split_chosen, idx_split_global, log_sis_ratio, logprior_nodeid, \
train_ids_left, train_ids_right, cache_tmp, loglik_left, loglik_right) \
= p.precomputed_proposal(data, param, settings, cache, node_id, train_ids, log_psplit)
if do_not_split_node_id:
p.do_not_split[node_id] = True
else:
p.update_left_right_statistics(cache_tmp, node_id, logprior_nodeid, train_ids_left,\
train_ids_right, loglik_left, loglik_right, feat_id_chosen, split_chosen, \
idx_split_global, settings, param, data, cache)
left, right = get_children_id(node_id)
grow_nodes.append(left)
grow_nodes.append(right)
# create mcmc structures
p.both_children_terminal.append(node_id)
parent = get_parent_id(node_id)
if (node_id != 0) and (parent in p.non_leaf_nodes):
p.inner_pc_pairs.append((parent, node_id))
if node_id != 0:
try:
p.both_children_terminal.remove(parent)
except ValueError:
pass
if settings.debug:
print 'sampled new tree:'
p.print_tree()
return p
def grow_next_pg(p, tree_pg, itr, settings):
p.log_sis_ratio = 0.
p.do_not_grow = False
p.grow_nodes = []
try:
nodes_processed = tree_pg.nodes_processed_itr[itr]
p.nodes_processed_itr.append(nodes_processed[:])
for node_id in nodes_processed[:-1]:
assert(tree_pg.do_not_split[node_id])
p.do_not_split[node_id] = True
node_id = nodes_processed[-1]
if node_id in tree_pg.node_info:
left, right = get_children_id(node_id)
log_sis_ratio_loglik_new = tree_pg.loglik[left] + tree_pg.loglik[right] - tree_pg.loglik[node_id]
try:
log_sis_ratio_loglik_old, log_sis_ratio_prior = tree_pg.log_sis_ratio_d[node_id]
except KeyError:
print 'tree_pg: node_info = %s, log_sis_ratio_d = %s' % (tree_pg.node_info, tree_pg.log_sis_ratio_d)
raise KeyError
if settings.verbose >= 2:
print 'log_sis_ratio_loglik_old = %s' % log_sis_ratio_loglik_old
print 'log_sis_ratio_loglik_new = %s' % log_sis_ratio_loglik_new
p.log_sis_ratio = log_sis_ratio_loglik_new + log_sis_ratio_prior
tree_pg.log_sis_ratio_d[node_id] = (log_sis_ratio_loglik_new, log_sis_ratio_prior)
p.log_sis_ratio_d[node_id] = tree_pg.log_sis_ratio_d[node_id]
p.non_leaf_nodes.append(node_id)
try:
p.leaf_nodes.remove(node_id)
except ValueError:
print 'warning: unable to remove node_id = %s from leaf_nodes = %s' % (node_id, p.leaf_nodes)
pass
p.leaf_nodes.append(left)
p.leaf_nodes.append(right)
# copying relevant bits
p.node_info[node_id] = tree_pg.node_info[node_id]
p.logprior[node_id] = tree_pg.logprior[node_id]
for node_id_child in [left, right]:
p.do_not_split[node_id_child] = False # can look up where node_id_child occurred in nodes_processed_itr
p.loglik[node_id_child] = tree_pg.loglik[node_id_child]
p.logprior[node_id_child] = tree_pg.logprior[node_id_child]
p.train_ids[node_id_child] = tree_pg.train_ids[node_id_child]
p.sum_y[node_id_child] = tree_pg.sum_y[node_id_child]
p.sum_y2[node_id_child] = tree_pg.sum_y2[node_id_child]
p.param_n[node_id_child] = tree_pg.param_n[node_id_child]
p.n_points[node_id_child] = tree_pg.n_points[node_id_child]
if settings.verbose >= 2:
print 'p.leaf_nodes = %s' % p.leaf_nodes
print 'p.non_leaf_nodes = %s' % p.non_leaf_nodes
print 'p.node_info.keys() = %s' % sorted(p.node_info.keys())
try:
p.grow_nodes = tree_pg.grow_nodes_itr[itr+1]
p.log_sis_ratio_d = tree_pg.log_sis_ratio_d
p.depth = tree_pg.depth
except IndexError:
p.do_not_grow = True
except IndexError:
p.do_not_grow = True
def grow_next_pg(p, tree_pg, itr, settings):
p.log_sis_ratio = 0.
p.do_not_grow = False
p.grow_nodes = []
try:
nodes_processed = tree_pg.nodes_processed_itr[itr]
p.nodes_processed_itr.append(nodes_processed[:])
for node_id in nodes_processed[:-1]:
assert(tree_pg.do_not_split[node_id])
p.do_not_split[node_id] = True
node_id = nodes_processed[-1]
if node_id in tree_pg.node_info:
left, right = get_children_id(node_id)
log_sis_ratio_loglik_new = tree_pg.loglik[left] + tree_pg.loglik[right] - tree_pg.loglik[node_id]
try:
log_sis_ratio_loglik_old, log_sis_ratio_prior = tree_pg.log_sis_ratio_d[node_id]
except KeyError:
print('tree_pg: node_info = %s, log_sis_ratio_d = %s' % (tree_pg.node_info, tree_pg.log_sis_ratio_d))
raise KeyError
if settings.verbose >= 2:
print('log_sis_ratio_loglik_old = %s' % log_sis_ratio_loglik_old)
print('log_sis_ratio_loglik_new = %s' % log_sis_ratio_loglik_new)
p.log_sis_ratio = log_sis_ratio_loglik_new + log_sis_ratio_prior
tree_pg.log_sis_ratio_d[node_id] = (log_sis_ratio_loglik_new, log_sis_ratio_prior)
p.log_sis_ratio_d[node_id] = tree_pg.log_sis_ratio_d[node_id]
p.non_leaf_nodes.append(node_id)
try:
p.leaf_nodes.remove(node_id)
except ValueError:
print('warning: unable to remove node_id = %s from leaf_nodes = %s' % (node_id, p.leaf_nodes))
pass
p.leaf_nodes.append(left)
p.leaf_nodes.append(right)
# copying relevant bits
p.node_info[node_id] = tree_pg.node_info[node_id]
p.logprior[node_id] = tree_pg.logprior[node_id]
for node_id_child in [left, right]:
p.do_not_split[node_id_child] = False # can look up where node_id_child occurred in nodes_processed_itr
p.loglik[node_id_child] = tree_pg.loglik[node_id_child]
p.logprior[node_id_child] = tree_pg.logprior[node_id_child]
p.train_ids[node_id_child] = tree_pg.train_ids[node_id_child]
p.sum_y[node_id_child] = tree_pg.sum_y[node_id_child]
p.sum_y2[node_id_child] = tree_pg.sum_y2[node_id_child]
p.param_n[node_id_child] = tree_pg.param_n[node_id_child]
p.n_points[node_id_child] = tree_pg.n_points[node_id_child]
if settings.verbose >= 2:
print('p.leaf_nodes = %s' % p.leaf_nodes)
print('p.non_leaf_nodes = %s' % p.non_leaf_nodes)
print('p.node_info.keys() = %s' % sorted(p.node_info.keys()))
try:
p.grow_nodes = tree_pg.grow_nodes_itr[itr+1]
p.log_sis_ratio_d = tree_pg.log_sis_ratio_d
p.depth = tree_pg.depth
except IndexError:
p.do_not_grow = True
except IndexError:
p.do_not_grow = True
def grow_next_pg(p,tree_pg,iterations,settings):
p.log_sis_ratio = 0.
p.do_not_grow = False
p.grow_nodes = []
try:
processed_nodes = tree_pg.processed_nodes_iterations[iterations]
p.processed_nodes_iterations.append(processed_nodes[:])
for id_of_node in processed_nodes[:-1]:
assert(tree_pg.do_not_split[id_of_node])
p.do_not_split[id_of_node] = True
id_of_node = processed_nodes[-1]
if id_of_node in tree_pg.node_info:
left_node,right_node = get_children_id(id_of_node)
log_sis_ratio_loglik_new = tree_pg.loglik[left_node] + tree_pg.loglik[right_node] - tree_pg.loglik[id_of_node]
try:
log_sis_ratio_loglik_old , log_sis_ratio_prior = tree_pg.log_sis_ratio_d[id_of_node]
except KeyError:
print('tree_pg: node_info =%s, log_sis_ratio_d = %s' % (tree_pg.node_info,tree_pg.log_sis_ratio_d))
raise KeyError
if settings.verbose >= 2:
print('log_sis_ratio_loglik_old =%s' % log_sis_ratio_loglik_old)
print('log_sis_ratio_loglik_new = %s' % log_sis_ratio_loglik_new)
p.log_sis_ratio = log_sis_ratio_loglik_new + log_sis_ratio_prior
tree_pg.log_sis_ratio_d[id_of_node] = (log_sis_ratio_loglik_new,log_sis_ratio_prior)
p.log_sis_ratio_d[id_of_node] = tree_pg.log_sis_ratio_d[id_of_node]
p.non_leaf_nodes.append(id_of_node)
try:
p.leaf_nodes.remove(id_of_node)
except ValueError:
print('Warning !!! Unable to remove id_of_node =%s from leaf_nodes = %s' %(id_of_node,p.leaf_nodes))
pass
p.leaf_nodes.append(left_node)
p.leaf_nodes.append(right_node)
p.node_info[id_of_node] = tree_pg.node_info[id_of_node]
p.logprior[id_of_node] = tree_pg.logprior[id_of_node]
for child_node_id in [left_node,right_node]:
p.do_not_split[child_node_id] = False
p.loglik[child_node_id] = tree_pg.loglik[child_node_id]
p.logprior[child_node_id] = tree_pg.logprior[child_node_id]
p.train_ids[child_node_id] = tree_pg.train_ids[child_node_id]
p.sum_y[child_node_id] = tree_pg.sum_y[child_node_id]
p.sum_y2[child_node_id] = tree_pg.sum_y2[child_node_id]
p.param_n[child_node_id] = tree_pg.param_n[child_node_id]
p.n_points[child_node_id] = tree_pg.n_points[child_node_id]
if settings.verbose >= 2:
print('p.leaf_nodes = %s' % p.leaf_nodes)
print('p.non_leaf_nodes =%s' % p.non_leaf_nodes)
print('p.node_info.keys() = %s' % sorted(p.node_info.keys()))
try:
p.grow_nodes = tree_pg.grow_nodes_iterations[iterations+1]
p.log_sis_ratio_d = tree_pg.log_sis_ratio_d
p.depth = tree_pg.depth
except IndexError:
p.do_not_grow = True
except IndexError:
p.do_not_grow = True
def get_nodes_subtree(self, node_id):
# NOTE: current node_id is included in nodes_subtree as well
node_list = []
expand = [node_id]
while len(expand) > 0:
node = expand.pop(0)
node_list.append(node)
if node not in self.leaf_nodes:
left, right = get_children_id(node)
expand.append(left)
expand.append(right)
return node_list
def get_nodes_subtree(self, node_id):
# NOTE: current node_id is included in nodes_subtree as well
node_list = []
expand = [node_id]
while len(expand) > 0:
node = expand.pop(0)
node_list.append(node)
if node not in self.leaf_nodes:
left, right = get_children_id(node)
expand.append(left)
expand.append(right)
return node_list
def evaluate_new_subtree(self, data, node_id_start, param, nodes_subtree,
cache, settings):
for i in self.train_ids[node_id_start]:
x_, y_ = data['x_train'][i, :], data['y_train'][i]
node_id = copy(node_id_start)
while True:
self.sum_y_new[node_id] += y_
self.sum_y2_new[node_id] += y_**2
self.n_points_new[node_id] += 1
self.train_ids_new[node_id] = np.append(
self.train_ids_new[node_id], i)
if node_id in self.leaf_nodes:
break
left, right = get_children_id(node_id)
feat_id, split, idx_split_global = self.node_info_new[
node_id] # splitting on new criteria
if x_[feat_id] <= split:
node_id = left
else:
node_id = right
for node_id in nodes_subtree:
self.loglik_new[node_id] = -np.inf
if self.n_points_new[node_id] > 0:
self.loglik_new[node_id], self.param_n_new[node_id] = \
compute_normal_normalizer(self.sum_y_new[node_id], self.sum_y2_new[node_id], \
self.n_points_new[node_id], param, cache, settings)
if node_id in self.leaf_nodes:
if stop_split(self.train_ids_new[node_id], settings, data,
cache):
# if leaf is empty, logprior_new[node_id] = 0.0 is incorrect; however
# loglik_new[node_id] = -np.inf will reject move to a tree with empty leaves
self.logprior_new[node_id] = 0.0
else:
# node with just 1 data point earlier could have more data points now
self.logprior_new[node_id] = np.log(
self.compute_pnosplit(node_id, param))
else:
# split probability might have changed if train_ids have changed
self.recompute_prob_split(data, param, settings, cache,
node_id)
if settings.debug == 1:
try:
check_if_zero(self.loglik[node_id_start] -
self.loglik_new[node_id_start])
except AssertionError:
print('train_ids[node_id_start] = %s, train_ids_new[node_id_start] = %s' \
% (self.train_ids[node_id_start], self.train_ids_new[node_id_start]))
raise AssertionError
def process_node_id(self,data,param , settings, cache, id_of_node):
if self.do_not_split[id_of_node]:
log_sis_ratio =0.0
else:
log_psplit = np.log(self.compute_psplit(id_of_node,param))
train_ids = self.train_ids[id_of_node]
left_node,right_node = get_children_id(id_of_node)
if setings.verbose >= 4:
print('Train_ids for this node =%s' % train_ids)
(dont_split_node_id,id_feat_chosen,chosen_split,split_global_idx,log_sis_ratio,logprior_nodeid,training_ids_left,training_ids_right,cache_tmp,loglikelihood_left,loglikelihood_right) =\
self.prior_proposal(data,param,settings,cache,id_of_node,train_ids,log_psplit)
if dont_split_node_id:
self.do_not_split[id_of_node] = True
else:
self.update_left_right_statistics(cache_tmp,id_of_node,logprior_nodeid,training_ids_left,training_ids_right,loglikelihood_left,loglikelihood_right,id_feat_chosen,chosen_split,split_global_idx,settings,param,data,cache)
self.grow_nodes.append(left_node)
self.grow_nodes.append(right_node)
return (log_sis_ratio)
def evaluate_new_subtree(self, data, node_id_start, param, nodes_subtree, cache, settings):
for i in self.train_ids[node_id_start]:
x_, y_ = data['x_train'][i, :], data['y_train'][i]
node_id = copy(node_id_start)
while True:
self.sum_y_new[node_id] += y_
self.sum_y2_new[node_id] += y_ ** 2
self.n_points_new[node_id] += 1
self.train_ids_new[node_id] = np.append(self.train_ids_new[node_id], i)
if node_id in self.leaf_nodes:
break
left, right = get_children_id(node_id)
feat_id, split, idx_split_global = self.node_info_new[node_id] # splitting on new criteria
if x_[feat_id] <= split:
node_id = left
else:
node_id = right
for node_id in nodes_subtree:
self.loglik_new[node_id] = -np.inf
if self.n_points_new[node_id] > 0:
self.loglik_new[node_id], self.param_n_new[node_id] = \
compute_normal_normalizer(self.sum_y_new[node_id], self.sum_y2_new[node_id], \
self.n_points_new[node_id], param, cache, settings)
if node_id in self.leaf_nodes:
if stop_split(self.train_ids_new[node_id], settings, data, cache):
# if leaf is empty, logprior_new[node_id] = 0.0 is incorrect; however
# loglik_new[node_id] = -np.inf will reject move to a tree with empty leaves
self.logprior_new[node_id] = 0.0
else:
# node with just 1 data point earlier could have more data points now
self.logprior_new[node_id] = np.log(self.compute_pnosplit(node_id, param))
else:
# split probability might have changed if train_ids have changed
self.recompute_prob_split(data, param, settings, cache, node_id)
if settings.debug == 1:
try:
check_if_zero(self.loglik[node_id_start] - self.loglik_new[node_id_start])
except AssertionError:
print 'train_ids[node_id_start] = %s, train_ids_new[node_id_start] = %s' \
% (self.train_ids[node_id_start], self.train_ids_new[node_id_start])
raise AssertionError
def compute_log_acc_g(self, node_id, param, len_both_children_terminal, loglik, \
train_ids_left, train_ids_right, cache, settings, data, grow_nodes):
# effect of do_not_split does not matter for node_id since it has children
logprior_children = 0.0
left, right = get_children_id(node_id)
if not no_valid_split_exists(data, cache, train_ids_left, settings):
logprior_children += np.log(self.compute_pnosplit(left, param))
if not no_valid_split_exists(data, cache, train_ids_right, settings):
logprior_children += np.log(self.compute_pnosplit(right, param))
log_acc_prior = np.log(self.compute_psplit(node_id, param)) \
-np.log(self.compute_pnosplit(node_id, param)) \
-np.log(len_both_children_terminal) + np.log(len(grow_nodes)) \
+ logprior_children
log_acc_loglik = (loglik - self.loglik[node_id])
log_acc = log_acc_prior + log_acc_loglik
if settings.verbose >= 2:
print 'compute_log_acc_g: log_acc_loglik = %s, log_acc_prior = %s' \
% (log_acc_loglik, log_acc_prior)
if loglik == -np.inf: # just need to ensure that an invalid split is not grown
log_acc = -np.inf
return log_acc
def process_node_id(self, data, param, settings, cache, node_id):
if self.do_not_split[node_id]:
log_sis_ratio = 0.0
else:
log_psplit = np.log(self.compute_psplit(node_id, param))
train_ids = self.train_ids[node_id]
left, right = get_children_id(node_id)
if settings.verbose >= 4:
print 'train_ids for this node = %s' % train_ids
(do_not_split_node_id, feat_id_chosen, split_chosen, idx_split_global, log_sis_ratio, logprior_nodeid, \
train_ids_left, train_ids_right, cache_tmp, loglik_left, loglik_right) \
= self.prior_proposal(data, param, settings, cache, node_id, train_ids, log_psplit)
if do_not_split_node_id:
self.do_not_split[node_id] = True
else:
self.update_left_right_statistics(cache_tmp, node_id, logprior_nodeid, train_ids_left,\
train_ids_right, loglik_left, loglik_right, feat_id_chosen, split_chosen, \
idx_split_global, settings, param, data, cache)
self.grow_nodes.append(left)
self.grow_nodes.append(right)
return (log_sis_ratio)
def process_node_id(self, data, param, settings, cache, node_id):
if self.do_not_split[node_id]:
log_sis_ratio = 0.0
else:
log_psplit = np.log(self.compute_psplit(node_id, param))
train_ids = self.train_ids[node_id]
left, right = get_children_id(node_id)
if settings.verbose >= 4:
print('train_ids for this node = %s' % train_ids)
(do_not_split_node_id, feat_id_chosen, split_chosen, idx_split_global, log_sis_ratio, logprior_nodeid, \
train_ids_left, train_ids_right, cache_tmp, loglik_left, loglik_right) \
= self.prior_proposal(data, param, settings, cache, node_id, train_ids, log_psplit)
if do_not_split_node_id:
self.do_not_split[node_id] = True
else:
self.update_left_right_statistics(cache_tmp, node_id, logprior_nodeid, train_ids_left,\
train_ids_right, loglik_left, loglik_right, feat_id_chosen, split_chosen, \
idx_split_global, settings, param, data, cache)
self.grow_nodes.append(left)
self.grow_nodes.append(right)
return (log_sis_ratio)
def compute_log_acc_g(self, node_id, param, len_both_children_terminal, loglik, \
train_ids_left, train_ids_right, cache, settings, data, grow_nodes):
# effect of do_not_split does not matter for node_id since it has children
logprior_children = 0.0
left, right = get_children_id(node_id)
if not no_valid_split_exists(data, cache, train_ids_left, settings):
logprior_children += np.log(self.compute_pnosplit(left, param))
if not no_valid_split_exists(data, cache, train_ids_right, settings):
logprior_children += np.log(self.compute_pnosplit(right, param))
log_acc_prior = np.log(self.compute_psplit(node_id, param)) \
-np.log(self.compute_pnosplit(node_id, param)) \
-np.log(len_both_children_terminal) + np.log(len(grow_nodes)) \
+ logprior_children
log_acc_loglik = (loglik - self.loglik[node_id])
log_acc = log_acc_prior + log_acc_loglik
if settings.verbose >= 2:
print('compute_log_acc_g: log_acc_loglik = %s, log_acc_prior = %s' \
% (log_acc_loglik, log_acc_prior))
if loglik == -np.inf: # just need to ensure that an invalid split is not grown
log_acc = -np.inf
return log_acc
def sample(self, data, settings, param, cache):
if settings.mcmc_type == 'growprune':
step_id = random.randint(0, 1) # only grow and prune moves permitted
elif settings.mcmc_type == 'cgm':
step_id = random.randint(0, 3) # all 4 moves equally likely (or think of 50% grow/prune, 25% change, 25% swap)
else:
raise Exception('invalid mcmc_type')
log_acc = -np.inf
log_r = 0
self.grow_nodes = [n_id for n_id in self.leaf_nodes \
if not stop_split(self.train_ids[n_id], settings, data, cache)]
grow_nodes = self.grow_nodes
if step_id == 0: # GROW step
if not grow_nodes:
change = False
else:
node_id = random.choice(grow_nodes)
if settings.verbose >= 1:
print 'grow_nodes = %s, chosen node_id = %s' % (grow_nodes, node_id)
do_not_split_node_id, feat_id, split, idx_split_global, logprior_nodeid = \
self.sample_split_prior(data, param, settings, cache, node_id)
assert not do_not_split_node_id
if settings.verbose >= 1:
print 'grow: do_not_split = %s, feat_id = %s, split = %s' \
% (do_not_split_node_id, feat_id, split)
train_ids = self.train_ids[node_id]
(train_ids_left, train_ids_right, cache_tmp, loglik_left, loglik_right) = \
compute_left_right_statistics(data, param, cache, train_ids, \
feat_id, split, settings)
loglik = loglik_left + loglik_right
len_both_children_terminal_new = len(self.both_children_terminal)
if get_sibling_id(node_id) not in self.leaf_nodes:
len_both_children_terminal_new += 1
log_acc = self.compute_log_acc_g(node_id, param, len_both_children_terminal_new, \
loglik, train_ids_left, train_ids_right, cache, settings, data, grow_nodes)
log_r = np.log(np.random.rand(1))
if log_r <= log_acc:
self.update_left_right_statistics(cache_tmp, node_id, logprior_nodeid, \
train_ids_left, train_ids_right, loglik_left, loglik_right, \
feat_id, split, idx_split_global, settings, param, data, cache)
# MCMC specific data structure updates
self.both_children_terminal.append(node_id)
parent = get_parent_id(node_id)
if (node_id != 0) and (parent in self.non_leaf_nodes):
self.inner_pc_pairs.append((parent, node_id))
sibling = get_sibling_id(node_id)
if sibling in self.leaf_nodes:
self.both_children_terminal.remove(parent)
change = True
else:
change = False
elif step_id == 1: # PRUNE step
if not self.both_children_terminal:
change = False # nothing to prune here
else:
node_id = random.choice(self.both_children_terminal)
feat_id = self.node_info[node_id][0]
if settings.verbose >= 1:
print 'prune: node_id = %s, feat_id = %s' % (node_id, feat_id)
left, right = get_children_id(node_id)
loglik = self.loglik[left] + self.loglik[right]
len_both_children_new = len(self.both_children_terminal)
grow_nodes_tmp = grow_nodes[:]
grow_nodes_tmp.append(node_id)
try:
grow_nodes_tmp.remove(left)
except ValueError:
pass
try:
grow_nodes_tmp.remove(right)
except ValueError:
pass
log_acc = - self.compute_log_inv_acc_p(node_id, param, len_both_children_new, \
loglik, grow_nodes_tmp, cache, settings, data)
log_r = np.log(np.random.rand(1))
if log_r <= log_acc:
self.remove_leaf_node_statistics(left, settings)
self.remove_leaf_node_statistics(right, settings)
self.leaf_nodes.append(node_id)
self.non_leaf_nodes.remove(node_id)
self.logprior[node_id] = np.log(self.compute_pnosplit(node_id, param))
# OK to set logprior as above since we know that a valid split exists
# MCMC specific data structure updates
self.both_children_terminal.remove(node_id)
parent = get_parent_id(node_id)
if (node_id != 0) and (parent in self.non_leaf_nodes):
self.inner_pc_pairs.remove((parent, node_id))
if node_id != 0:
sibling = get_sibling_id(node_id)
if sibling in self.leaf_nodes:
if settings.debug == 1:
assert(parent not in self.both_children_terminal)
self.both_children_terminal.append(parent)
change = True
else:
change = False
elif step_id == 2: # CHANGE
if not self.non_leaf_nodes:
change = False
else:
node_id = random.choice(self.non_leaf_nodes)
do_not_split_node_id, feat_id, split, idx_split_global, logprior_nodeid = \
self.sample_split_prior(data, param, settings, cache, node_id)
if settings.verbose >= 1:
print 'change: node_id = %s, do_not_split = %s, feat_id = %s, split = %s' \
% (node_id, do_not_split_node_id, feat_id, split)
# Note: this just samples a split criterion, not guaranteed to "change"
assert(not do_not_split_node_id)
nodes_subtree = self.get_nodes_subtree(node_id)
nodes_not_in_subtree = self.get_nodes_not_in_subtree(node_id)
if settings.debug == 1:
set1 = set(list(nodes_subtree) + list(nodes_not_in_subtree))
set2 = set(self.leaf_nodes + self.non_leaf_nodes)
assert(sorted(set1) == sorted(set2))
self.create_new_statistics(nodes_subtree, nodes_not_in_subtree, node_id, settings)
self.node_info_new[node_id] = (feat_id, split, idx_split_global)
self.evaluate_new_subtree(data, node_id, param, nodes_subtree, cache, settings)
# log_acc will be be modified below
log_acc_tmp, loglik_diff, logprior_diff = self.compute_log_acc_cs(nodes_subtree, node_id)
if settings.debug == 1:
self.check_if_same(log_acc_tmp, loglik_diff, logprior_diff)
log_acc = log_acc_tmp + self.logprior[node_id] - self.logprior_new[node_id]
log_r = np.log(np.random.rand(1))
if log_r <= log_acc:
self.node_info[node_id] = copy(self.node_info_new[node_id])
self.update_subtree(node_id, nodes_subtree, settings)
change = True
else:
change = False
elif step_id == 3: # SWAP
if not self.inner_pc_pairs:
change = False
else:
node_id, child_id = random.choice(self.inner_pc_pairs)
nodes_subtree = self.get_nodes_subtree(node_id)
nodes_not_in_subtree = self.get_nodes_not_in_subtree(node_id)
if settings.debug == 1:
set1 = set(list(nodes_subtree) + list(nodes_not_in_subtree))
set2 = set(self.leaf_nodes + self.non_leaf_nodes)
assert(sorted(set1) == sorted(set2))
self.create_new_statistics(nodes_subtree, nodes_not_in_subtree, node_id, settings)
self.node_info_new[node_id] = copy(self.node_info[child_id])
self.node_info_new[child_id] = copy(self.node_info[node_id])
if settings.verbose >= 1:
print 'swap: node_id = %s, child_id = %s' % (node_id, child_id)
print 'node_info[node_id] = %s, node_info[child_id] = %s' \
% (self.node_info[node_id], self.node_info[child_id])
self.evaluate_new_subtree(data, node_id, param, nodes_subtree, cache, settings)
log_acc, loglik_diff, logprior_diff = self.compute_log_acc_cs(nodes_subtree, node_id)
if settings.debug == 1:
self.check_if_same(log_acc, loglik_diff, logprior_diff)
log_r = np.log(np.random.rand(1))
if log_r <= log_acc:
self.node_info[node_id] = copy(self.node_info_new[node_id])
self.node_info[child_id] = copy(self.node_info_new[child_id])
self.update_subtree(node_id, nodes_subtree, settings)
change = True
else:
change = False
if settings.verbose >= 1:
print 'trying move: step_id = %d, move = %s, log_acc = %s, log_r = %s' \
% (step_id, STEP_NAMES[step_id], log_acc, log_r)
if change:
self.depth = max([get_depth(node_id) for node_id in \
self.leaf_nodes])
self.loglik_current = sum([self.loglik[node_id] for node_id in \
self.leaf_nodes])
if settings.verbose >= 1:
print 'accepted move: step_id = %d, move = %s' % (step_id, STEP_NAMES[step_id])
self.print_stuff()
if settings.debug == 1:
both_children_terminal, inner_pc_pairs = self.recompute_mcmc_data_structures()
print '\nstats from recompute_mcmc_data_structures'
print 'both_children_terminal = %s' % both_children_terminal
print 'inner_pc_pairs = %s' % inner_pc_pairs
assert(sorted(both_children_terminal) == sorted(self.both_children_terminal))
assert(sorted(inner_pc_pairs) == sorted(self.inner_pc_pairs))
grow_nodes_new = [n_id for n_id in self.leaf_nodes \
if not stop_split(self.train_ids[n_id], settings, data, cache)]
if change and (step_id == 1):
print 'grow_nodes_new = %s, grow_nodes_tmp = %s' % (sorted(grow_nodes_new), sorted(grow_nodes_tmp))
assert(sorted(grow_nodes_new) == sorted(grow_nodes_tmp))
return (change, step_id)
def sample(self, data, settings, param, cache,databc,cachebc):
if settings.mcmc_type == 'growprune':
step_id = random.randint(0, 1) # only grow and prune moves permitted
elif settings.mcmc_type == 'cgm':
step_id = random_pick([0,1,2,3], [0.25, 0.25, 0.4, 0.1])
else:
raise Exception('invalid mcmc_type')
log_acc = -np.inf
log_r = 0
self.grow_nodes = [n_id for n_id in self.leaf_nodes \
if not stop_split(self.train_ids[n_id], settings, data, cache)]
grow_nodes = self.grow_nodes
if step_id == 0:
if not grow_nodes:
change = False
else:
node_id = random.choice(grow_nodes)
if settings.verbose >= 1:
print 'grow_nodes = %s, chosen node_id = %s' % (grow_nodes, node_id)
do_not_split_node_id, feat_id, split, idx_split_global, logprior_nodeid = \
self.sample_split_prior(data, param, settings, cache, node_id,databc,cachebc)
if do_not_split_node_id==True:
change=False
if do_not_split_node_id==False:
if settings.verbose >= 1:
print 'grow: do_not_split = %s, feat_id = %s, split = %s' \
% (do_not_split_node_id, feat_id, split)
train_ids = self.train_ids[node_id]
t0 = time.time()
if settings.parallelize_compute_statistics==0:
(train_ids_left, train_ids_right, cache_tmp, loglik_left, loglik_right) = \
compute_left_right_statistics(data, param, cache, train_ids, feat_id, split, settings)
else:
(cache_tmp, loglik_left, loglik_right) = \
compute_left_right_statistics_parallel(data, param, cache, train_ids, feat_id, split, settings, databc)
cond = data['x_train'][train_ids, feat_id] <= split
train_ids_left = train_ids[cond]
train_ids_right = train_ids[~cond]
if settings.timer==1:
settings.parallelize_compute_statistics_time += time.time()-t0
loglik = loglik_left + loglik_right
len_both_children_terminal_new = len(self.both_children_terminal)
if get_sibling_id(node_id) not in self.leaf_nodes:
len_both_children_terminal_new += 1
log_acc = self.compute_log_acc_g(node_id, param, len_both_children_terminal_new, \
loglik, train_ids_left, train_ids_right, cache, settings, data, grow_nodes)
log_r = np.log(np.random.rand(1))
if log_r <= log_acc:
self.update_left_right_statistics(cache_tmp, node_id, logprior_nodeid, \
train_ids_left, train_ids_right, loglik_left, loglik_right, \
feat_id, split, idx_split_global, settings, param, data, cache)
# MCMC specific data structure updates
self.both_children_terminal.append(node_id)
parent = get_parent_id(node_id)
if (node_id != 0) and (parent in self.non_leaf_nodes):
self.inner_pc_pairs.append((parent, node_id))
sibling = get_sibling_id(node_id)
if sibling in self.leaf_nodes:
self.both_children_terminal.remove(parent)
change = True
else:
change = False
elif step_id == 1:
if not self.both_children_terminal:
change = False
else:
node_id = random.choice(self.both_children_terminal)
feat_id = self.node_info[node_id][0]
if settings.verbose >= 1:
print 'prune: node_id = %s, feat_id = %s' % (node_id, feat_id)
left, right = get_children_id(node_id)
loglik = self.loglik[left] + self.loglik[right]
len_both_children_new = len(self.both_children_terminal)
grow_nodes_tmp = grow_nodes[:]
grow_nodes_tmp.append(node_id)
try:
grow_nodes_tmp.remove(left)
except ValueError:
pass
try:
grow_nodes_tmp.remove(right)
except ValueError:
pass
log_acc = - self.compute_log_inv_acc_p(node_id, param, len_both_children_new, \
loglik, grow_nodes_tmp, cache, settings, data)
log_r = np.log(np.random.rand(1))
if log_r <= log_acc:
self.remove_leaf_node_statistics(left, settings)
self.remove_leaf_node_statistics(right, settings)
self.leaf_nodes.append(node_id)
self.non_leaf_nodes.remove(node_id)
self.logprior[node_id] = np.log(self.compute_pnosplit(node_id, param))
# OK to set logprior as above since we know that a valid split exists
# MCMC specific data structure updates
self.both_children_terminal.remove(node_id)
parent = get_parent_id(node_id)
if (node_id != 0) and (parent in self.non_leaf_nodes):
self.inner_pc_pairs.remove((parent, node_id))
if node_id != 0:
sibling = get_sibling_id(node_id)
if sibling in self.leaf_nodes:
if settings.debug == 1:
assert(parent not in self.both_children_terminal)
self.both_children_terminal.append(parent)
change = True
else:
change = False
elif step_id == 2:
if not self.non_leaf_nodes:
change = False
else:
node_id = random.choice(self.non_leaf_nodes)
do_not_split_node_id, feat_id, split, idx_split_global, logprior_nodeid = \
self.sample_split_prior(data, param, settings, cache, node_id,databc,cachebc)
if settings.verbose >= 1:
print 'change: node_id = %s, do_not_split = %s, feat_id = %s, split = %s' \
% (node_id, do_not_split_node_id, feat_id, split)
# Note: this just samples a split criterion, not guaranteed to "change"
#assert(not do_not_split_node_id)
if do_not_split_node_id==True:
change = False
if do_not_split_node_id==False:
nodes_subtree = self.get_nodes_subtree(node_id)
nodes_not_in_subtree = self.get_nodes_not_in_subtree(node_id)
if settings.debug == 1:
set1 = set(list(nodes_subtree) + list(nodes_not_in_subtree))
set2 = set(self.leaf_nodes + self.non_leaf_nodes)
assert(sorted(set1) == sorted(set2))
self.create_new_statistics(nodes_subtree, nodes_not_in_subtree, node_id, settings)
self.node_info_new[node_id] = (feat_id, split, idx_split_global)
self.evaluate_new_subtree(data, node_id, param, nodes_subtree, cache, settings, databc, cachebc)
# log_acc will be be modified below
log_acc_tmp, loglik_diff, logprior_diff = self.compute_log_acc_cs(nodes_subtree, node_id)
if settings.debug == 1:
self.check_if_same(log_acc_tmp, loglik_diff, logprior_diff)
log_acc = log_acc_tmp + self.logprior[node_id] - self.logprior_new[node_id]
log_r = np.log(np.random.rand(1))
if log_r <= log_acc:
self.node_info[node_id] = copy(self.node_info_new[node_id])
self.update_subtree(node_id, nodes_subtree, settings)
change = True
else:
change = False
elif step_id == 3:
if not self.inner_pc_pairs:
change = False
else:
node_id, child_id = random.choice(self.inner_pc_pairs)
nodes_subtree = self.get_nodes_subtree(node_id)
nodes_not_in_subtree = self.get_nodes_not_in_subtree(node_id)
if settings.debug == 1:
set1 = set(list(nodes_subtree) + list(nodes_not_in_subtree))
set2 = set(self.leaf_nodes + self.non_leaf_nodes)
assert(sorted(set1) == sorted(set2))
self.create_new_statistics(nodes_subtree, nodes_not_in_subtree, node_id, settings)
self.node_info_new[node_id] = copy(self.node_info[child_id])
self.node_info_new[child_id] = copy(self.node_info[node_id])
if settings.verbose >= 1:
print 'swap: node_id = %s, child_id = %s' % (node_id, child_id)
print 'node_info[node_id] = %s, node_info[child_id] = %s' \
% (self.node_info[node_id], self.node_info[child_id])
self.evaluate_new_subtree(data, node_id, param, nodes_subtree, cache, settings, databc, cachebc)
log_acc, loglik_diff, logprior_diff = self.compute_log_acc_cs(nodes_subtree, node_id)
if settings.debug == 1:
self.check_if_same(log_acc, loglik_diff, logprior_diff)
log_r = np.log(np.random.rand(1))
if log_r <= log_acc:
self.node_info[node_id] = copy(self.node_info_new[node_id])
self.node_info[child_id] = copy(self.node_info_new[child_id])
self.update_subtree(node_id, nodes_subtree, settings)
change = True
else:
change = False
if settings.verbose >= 1:
print 'trying move: step_id = %d, move = %s, log_acc = %s, log_r = %s' \
% (step_id, STEP_NAMES[step_id], log_acc, log_r)
if change:
self.depth = max([get_depth(node_id) for node_id in \
self.leaf_nodes])
self.loglik_current = sum([self.loglik[node_id] for node_id in \
self.leaf_nodes])
if settings.verbose >= 1:
print 'accepted move: step_id = %d, move = %s' % (step_id, STEP_NAMES[step_id])
self.print_stuff()
if settings.debug == 1:
both_children_terminal, inner_pc_pairs = self.recompute_mcmc_data_structures()
print '\nstats from recompute_mcmc_data_structures'
print 'both_children_terminal = %s' % both_children_terminal
print 'inner_pc_pairs = %s' % inner_pc_pairs
assert(sorted(both_children_terminal) == sorted(self.both_children_terminal))
assert(sorted(inner_pc_pairs) == sorted(self.inner_pc_pairs))
grow_nodes_new = [n_id for n_id in self.leaf_nodes \
if not stop_split(self.train_ids[n_id], settings, data, cache)]
if change and (step_id == 1):
print 'grow_nodes_new = %s, grow_nodes_tmp = %s' % (sorted(grow_nodes_new), sorted(grow_nodes_tmp))
assert(sorted(grow_nodes_new) == sorted(grow_nodes_tmp))
return (change, step_id)
