File size: 4,271 Bytes
b5567db | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 | import sympy as sp
import numpy as np
from sklearn.metrics import mutual_info_score
# 符号
import sympy as sp
import pandas as pd
# symbols
X = sp.Symbol("X")
Y = sp.Symbol("Y")
Z = sp.Symbol("Z")
class MI(sp.Function):
nargs = (2,)
class CMI(sp.Function):
nargs = (3,)
class II(sp.Function):
nargs = (3,) # interaction information
ALLOWED_LOCALS = {
"X": X,
"Y": Y,
"Z": Z,
"I": MI, # I(X,Y)
"CI": CMI, # I(X,Y|Z) 条件互信息
"II": II # I(X；Y；Z)交互信息
}
def parse_expression(expr_str: str) -> sp.Expr:
"""
String → SymPy Expression
"""
expr = sp.sympify(expr_str, locals=ALLOWED_LOCALS)
return expr
def entropy(x):#计算熵
_, cnt = np.unique(x, return_counts=True)
p = cnt / cnt.sum()
return -np.sum(p * np.log(p + 1e-12))
def mi(x, y):#互信息
return mutual_info_score(x, y)
def cmi(x, y, z):#条件互信息（通过熵的加减计算）
# I(X;Y|Z) = H(X,Z)+H(Y,Z)-H(Z)-H(X,Y,Z)
return (
entropy(np.c_[x, z].tolist())
+ entropy(np.c_[y, z].tolist())
- entropy(z)
- entropy(np.c_[x, y, z].tolist())
)
def interaction_info(x, y, z):#交互信息
# I(X;Y;Z) = I(X;Y) - I(X;Y|Z)
return mi(x, y) - cmi(x, y, z)
def expr_to_callable(expr: sp.Expr):
def eval_node(node, ctx):
if isinstance(node, MI):
return mi(eval_node(node.args[0], ctx),
eval_node(node.args[1], ctx))
if isinstance(node, CMI):
return cmi(eval_node(node.args[0], ctx),
eval_node(node.args[1], ctx),
eval_node(node.args[2], ctx))
if isinstance(node, II):
return interaction_info(
eval_node(node.args[0], ctx),
eval_node(node.args[1], ctx),
eval_node(node.args[2], ctx)
)
if node == X:
return ctx["X"]
if node == Y:
return ctx["Y"]
if node == Z:
return ctx["Z"]
if node.is_Number:
return float(node)
if node.is_Add:
return sum(eval_node(arg, ctx) for arg in node.args)
if node.is_Mul:
r = 1.0
for arg in node.args:
r *= eval_node(arg, ctx)
return r
if node.is_Pow:
base, exp = node.args
return eval_node(base, ctx) ** eval_node(exp, ctx)
raise ValueError(f"Unsupported node: {node}")
def f(X_arr, Y_arr, Z_arr=None):
ctx = {"X": X_arr, "Y": Y_arr}
if Z_arr is not None:
ctx["Z"] = Z_arr
return eval_node(expr, ctx)
return f
from sklearn.preprocessing import LabelEncoder
def changetosinge(x):
return float(x)
# scores = f(X, y, X_other_list)
def prepare_data(dataname, base_url):
url = os.path.join(base_url, dataname + '.mat')
data = scio.loadmat(url)
X0 = pd.DataFrame(data['X'])
y0 = pd.DataFrame(data['Y'])
if dataname == 'Dermatology':
Special = X0.iloc[:, -1]
a = np.array([item[0] for item in Special])
label_encoder = LabelEncoder()
a33 = label_encoder.fit_transform(a)
X0 = X0.iloc[:, :-1]
X0[33] = a33
X0 = X0.applymap(changetosinge)
y0 = y0.applymap(changetosinge)
label_encoder = LabelEncoder()
y_encoded = label_encoder.fit_transform(y0)
y = pd.DataFrame(y_encoded)
X = pd.DataFrame()
for col in X0.columns:
X[col] = pd.cut(X0[col], bins=5, labels=False)
new_columns = [str(i) for i in range(X.shape[1] + 1)]
X = X.rename(columns=dict(zip(X.columns, new_columns[:-1])))
y = y.rename(columns=dict(zip(y.columns, [new_columns[-1]])))
data_processed = pd.concat([X, y], axis=1)
# data_processed = pd.DataFrame(X)
return data_processed, list(set(y_encoded))
import os
import scipy.io as scio
dataname = 'Authorship'
base_url = '/home/fangsensen/AutoFS/data/'
data_processed, class_set = prepare_data(dataname, base_url)
# print(data_processed)
# X_arr = data_processed['0']
# y_arr = data_processed['69']
print(111111,X_arr,2222222,y_arr)
expr = parse_expression("I(X,Y)")
f = expr_to_callable(expr)
score = f(X_arr, y_arr)
print(score)
|