def graph_cat(err_level, nominal_num, category_num):
G = nx.DiGraph()
for i in range(30):
# c stands for category choice
q = normal_distribution()
G.add_node(i, quality=q)
pair_list = all_pairs(30)
for pair in pair_list:
G.add_edge(pair[0], pair[1], weight=G.node[pair[1]]['quality'])
return G
def graph_nominal(err_level, nominal_num, category_num):
G = nx.DiGraph()
for i in range(30):
q = normal_distribution()
n = nominal_value(nominal_num)
G.add_node(i, quality=q, nominal=n)
pair_list = all_pairs(30)
for pair in pair_list:
attachment = G.node[pair[1]]['nominal']
G.add_edge(pair[0], pair[1], weight=attachment)
return G
def graph_ddd(err_level, nominal_num, category_num):
G = nx.DiGraph()
for i in range(30):
q = normal_distribution()
G.add_node(i, quality=q)
pair_list = all_pairs(30)
for pair in pair_list:
# initialize the weight as the sum of quality of the to_node and the error of the to_node
attachment = G.node[pair[1]]['quality']
G.add_edge(pair[0], pair[1], weight=attachment)
return G
def graph_collective_err(err_level, nominal_num, category_num):
"""
# Graph for collective level error network
# Error of the attachment is err~N(0, x), x is the error level
# At initialization, the quality and attachment are all independent randomized value
"""
G = nx.DiGraph()
for i in range(30):
q = normal_distribution()
e = error_func(err_level)
G.add_node(i, quality=q, error=e)
pair_list = all_pairs(30)
for pair in pair_list:
# initialize the weight as the sum of quality of the to_node and the error of the to_node
attachment = G.node[pair[1]]['quality'] + G.node[pair[1]]['error']
G.add_edge(pair[0], pair[1], weight=attachment)
return G
def graph_dyadic_err(err_level, nominal_num, category_num):
"""
Graph for dyadic level error network
The attachment between nodes are the sum of quality of the to_node and an error
Error of the attachment is err~N(0, x), x is the error level
At initialization, the quality and attachment are all independent randomized value
"""
G = nx.DiGraph()
for i in range(30):
q = normal_distribution()
G.add_node(i, quality=q)
pair_list = all_pairs(30)
for pair in pair_list:
# initialize the weight as the sum of quality of the to_node and an error
attachment = G.node[pair[1]]['quality'] + error_func(err_level)
G.add_edge(pair[0], pair[1], weight=attachment)
return G
def graph_rd(err_level, nominal_num, category_num):
"""
Graph for randomized network
The attachment between nodes are random value from N~(0, 1)
At initialization, the quality and attachment are all independent randomized value
"""
G = nx.DiGraph()
for i in range(30):
q = normal_distribution()
G.add_node(i, quality=q)
# pdb.set_trace()
pair_list = all_pairs(30)
for pair in pair_list:
# initialize the weight as random value that follows a normal distribution
attachment = normal_distribution()
G.add_edge(pair[0], pair[1], weight=attachment)
return G