-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmusic2.py
181 lines (164 loc) · 6.41 KB
/
music2.py
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
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
from utils.disp import scatter_data_3d,scatter_2data_3d
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from utils.k_fold_cross_validation import Khold
from utils.data_analize import correlation,to2D
from utils.ts_fuzzy.ts_fuzzy import TSFuzzy
from utils.ts_fuzzy.actfunc import ActiveFunc
from utils.optimize_algorithms.particle_swarm_optimization import PSO
import sklearn.metrics as metrics
from multiprocessing import Pool
from pylab import rcParams
import matplotlib.pyplot as plt
import seaborn as sns
import os
from line import send_message
M = 2.1
C = 5
Node = 5
ParMax = 5
Size = 100
TMax = 500
W = 0.9
CP = 0.8
CG = 0.6
def learn(input):
x_train = input[0]
y_train = input[1]
x_test = input[2]
y_test = input[3]
x_src = input[4]
y_src = input[5]
pso = PSO(size=Size,
tMax=TMax,
w=W,
cp=CP,
cg=CG,
parMax=ParMax,
X_src=x_src,
y_src=y_src,
X_tgt=x_train,
y_tgt=y_train,
X_test=x_test,
y_test=y_test,
M=M,
C=C,
Node=Node,
ActFunc=ActiveFunc.sigmoid)
gBestValue,y_pred,trans,gBestPos = pso.learn()
return [metrics.mean_squared_error(y_pred,y_test),metrics.r2_score(y_pred,y_test)]
def noTLLearn(input):
x_src = input[0]
y_src = input[1]
x_test = input[2]
y_test = input[3]
ts = TSFuzzy(x_src,y_src,None,None,M,C,Node,ParMax,ActiveFunc.sigmoid)
y_pred = ts.predict(x_test)
return [metrics.mean_squared_error(y_pred,y_test),metrics.r2_score(y_pred,y_test)]
def NoTL(x_tgt,y_tgt,dir):
if os.path.exists(f"{dir}/notl.csv"):
return
khold = Khold(x_tgt.T,y_tgt,10)
i = 0
result = []
arg_list = []
for x_train,y_train,x_test,y_test in khold:
arg_list.append([x_train,y_train,x_test,y_test])
with Pool(processes=6) as p:
result = p.map(func=noTLLearn,iterable=arg_list)
df = pd.DataFrame(result,columns=["mean squared error","R"])
df.to_csv(f"{dir}/notl.csv")
return np.mean(df["mean squared error"])
def ReduceFeatureTL(x_tgt,y_tgt,x_src,y_src,dir):
if os.path.exists(f"{dir}/reduce.csv"):
return
khold = Khold(x_tgt.T,y_tgt,10)
i = 0
result = []
arg_list = []
for x_train,y_train,x_test,y_test in khold:
arg_list.append([x_train,y_train,x_test,y_test,x_src,y_src])
with Pool(processes=6) as p:
result = p.map(func=learn,iterable=arg_list)
df = pd.DataFrame(result,columns=["mean squared error","R"])
df.to_csv(f"{dir}/reduce.csv")
return np.mean(df["mean squared error"])
def MappingFeatureTL(x_tgt,y_tgt,x_src,y_src,dir):
if os.path.exists(f"{dir}/mapping.csv"):
return
khold = Khold(x_tgt.T,y_tgt,10)
i = 0
result = []
arg_list = []
for x_train,y_train,x_test,y_test in khold:
arg_list.append([x_train,y_train,x_test,y_test,x_src,y_src])
with Pool(processes=6) as p:
result = p.map(func=learn,iterable=arg_list)
df = pd.DataFrame(result,columns=["mean squared error","R"])
df.to_csv(f"{dir}/mapping.csv")
return np.mean(df["mean squared error"])
def AllFeatureTL(x_tgt,y_tgt,x_src,y_src,dir):
khold = Khold(x_tgt.T,y_tgt,10)
result = []
arg_list = []
for x_train,y_train,x_test,y_test in khold:
arg_list.append([x_train,y_train,x_test,y_test,x_src,y_src])
with Pool(processes=6) as p:
result = p.map(func=learn,iterable=arg_list)
df = pd.DataFrame(result,columns=["mean squared error","R"])
df.to_csv(f"{dir}/all.csv")
return np.mean(df["mean squared error"])
def get_data(filename):
missing_values = ["n/a", "na", "--", " ", "N/A", "NA"]
return pd.read_csv(filename,na_values = missing_values,sep=",")
if __name__ == "__main__":
originalParty = get_data("data/proced_data/spotify/genre/party.csv")
src = originalParty.copy()
originalPunk = get_data("data/proced_data/spotify/genre/punk.csv")
tgt = originalPunk.copy()
columns = ['year','danceability','key','loudness','mode','speechiness','acousticness','instrumentalness','liveness','valence','tempo','duration_ms','time_signature']
no_tl_result = np.zeros([len(columns),len(columns)])
mapping_result = np.zeros([len(columns),len(columns)])
reduce_result = np.zeros([len(columns),len(columns)])
for src_idx in range(len(columns)):
for tgt_idx in range(len(columns)):
if(src_idx == tgt_idx):
continue
src_col = np.append(np.delete(columns,[src_idx,tgt_idx]),columns[src_idx])
tgt_col = np.append(np.delete(columns,[src_idx,tgt_idx]),columns[tgt_idx])
reduce_col = np.delete(columns,[src_idx,tgt_idx])
y_src = src['popularity']
y_tgt = tgt['popularity']
x_src = []
x_tgt = []
x_src_reduce = []
x_tgt_reduce = []
for col in src_col:
x_src.append(src[col])
for col in tgt_col:
x_tgt.append(tgt[col])
for col in reduce_col:
x_src_reduce.append(src[col])
x_tgt_reduce.append(tgt[col])
x_src = np.array(x_src)
x_tgt = np.array(x_tgt)
x_src_reduce = np.array(x_src_reduce)
x_tgt_reduce = np.array(x_tgt_reduce)
dir = f"result/music/{columns[src_idx]}_to_{columns[tgt_idx]}"
print((src_idx)*len(columns) + (tgt_idx + 1),"/",len(columns)*len(columns)," ",dir)
if not os.path.exists(dir):
send_message(str((src_idx)*len(columns) + (tgt_idx + 1))+"/"+str(len(columns)*len(columns))+" "+dir)
os.mkdir(dir)
print("no tl")
no_tl_result[src_idx][tgt_idx] = NoTL(x_tgt,y_tgt,dir)
print("reduce")
reduce_result[src_idx][tgt_idx] = ReduceFeatureTL(x_tgt_reduce,y_tgt,x_src_reduce,y_src,dir)
print("mapping")
mapping_result[src_idx][tgt_idx] = MappingFeatureTL(x_tgt,y_tgt,x_src,y_src,dir)
resultDF = pd.DataFrame(no_tl_result,columns = columns)
resultDF.to_excel("result/music/no_tl_result.xlsx")
resultDF = pd.DataFrame(reduce_result,columns = columns)
resultDF.to_excel("result/music/reduce_result.xlsx")
resultDF = pd.DataFrame(mapping_result,columns = columns)
resultDF.to_excel("result/music/mapping_result.xlsx")