-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpredict.py
205 lines (189 loc) · 9.08 KB
/
predict.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
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
# モジュールのimport
from pytrends.request import TrendReq
from itertools import zip_longest
from itertools import filterfalse
import pandas as pd
import csv
import pprint
# 学習用モジュールインポート
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
pytrends = TrendReq(hl='ja-JP', tz=360)
data = pd.read_excel('2016.xlsx')
df = pd.DataFrame(data["キーワードリスト"])
value_list = df.values.tolist()
list3 = []
for list1 in value_list:
list2 = list1[0].split(',')
list3.append(list2)
results = []
group_by = 4
for i in range(len(list3)):
primary_kw = list3[i][0]
del list3[i][0]
kw_list = list3[i]
chunks = zip_longest(*[iter(kw_list)]*group_by)
p = lambda x: x is None
merged_df = None
for elems in list(chunks):
elems = list(filterfalse(p, elems))
elems.append(primary_kw)
pytrends.build_payload(elems, cat=0, timeframe='2016-07-01 2016-07-10', geo='JP', gprop='')
df = pytrends.interest_over_time()
# 取得結果のスコアは String になる。 float 変換したいので、True/False が設定されている`isPartial` を削除して float に変換する。
del df['isPartial']
df = df.astype('float64')
# dataframe を primary_kw で最大値で正規化する
scaled_dataframe = df.div(df[primary_kw].max(), axis=0)
if merged_df is None:
merged_df = scaled_dataframe
else:
# ValueError: columns overlap but no suffix specified が発生するので、やむなく、'postgresql' を削除
del scaled_dataframe[primary_kw]
merged_df = merged_df.join(scaled_dataframe)
results.append(merged_df.mean()/merged_df.mean().mean())
res = pd.DataFrame(pd.DataFrame(results).sum())
res.columns = [ "trend"]
res["name"] = res.index
res.name = res.name.str.replace("'", "")
res = res.reset_index(drop = True)
res.head()
election = pd.read_csv("2016_cand_ver2.csv")
election.name = election.name.str.replace(" ", "")
#election.status = election.where(election.status >= 2, 1)
election = election.drop(["num_votes", "vote_ratio"], axis = 1)
election.sex = (election.sex == "男")
# 党の整理
election.party = election.party.str.replace("おおさか維新の会", "維新").replace("維新政党新風", "維新")
election.party = election.party.str.replace("こころ", "s").replace("支持政党なし",
"s").replace("国民怒りの声", "s").replace("新党改革","s").replace("世界経済共同体党",
"s").replace("犬丸勝子と共和党", "s").replace("地球平和党", "s").replace("チャレンジド日本",
"s").replace("減税日本", "s").replace("S", "s")
pd.unique(election.party)
# 検索トレンドと候補者データのmerge
final = pd.merge(election, res, on = "name", how = "left")
final= final.fillna(0)
final.head()
# 学習用データ（説明変数）の作成
party = pd.get_dummies(final["party"])
party.head()
X = pd.merge(final.drop(['name', 'party','elected', 'area'],axis=1), party, left_index=True, right_index=True)
X.head()
# target
Y = final.elected
# ロジスティック回帰モデルのインスタンスを作成
lr = LogisticRegression()
# ロジスティック回帰モデルの重みを学習
lr.fit(X, Y)
#データの読み込み
data_19 = pd.read_csv("candidates_19.csv")
data_19.head()
# 党の整理
data_19.party = data_19.party.str.replace("立憲", "民進").replace("国民", "民進")
data_19.party = data_19.party.str.replace("安楽", "s").replace("労働者",
"s").replace("町田", "s").replace("高橋","s").replace("N国",
"s").replace("加藤", "s").replace("オリーブ", "s").replace("れいわ",
"s").replace("日本無党派党", "s")
data_19.head()
data = pd.read_excel('2019.xlsx')
df = pd.DataFrame(data["キーワードリスト"])
value_list = df.values.tolist()
list3 = []
for list1 in value_list:
list2 = list1[0].split(',')
list3.append(list2)
results = []
group_by = 4
for i in range(len(list3)):
primary_kw = list3[i][0]
del list3[i][0]
kw_list = list3[i]
chunks = zip_longest(*[iter(kw_list)]*group_by)
p = lambda x: x is None
merged_df = None
for elems in list(chunks):
elems = list(filterfalse(p, elems))
elems.append(primary_kw)
pytrends.build_payload(elems, cat=0, timeframe='2019-07-10 2019-07-19', geo='JP', gprop='')
df = pytrends.interest_over_time()
# 取得結果のスコアは String になる。 float 変換したいので、True/False が設定されている`isPartial` を削除して float に変換する。
# if 'isPartial' in df.columns:
if df.empty is not True:
del df['isPartial']
df = df.astype('float64')
# dataframe を primary_kw で最大値で正規化する
scaled_dataframe = df.div(df[primary_kw].max(), axis=0)
if merged_df is None:
merged_df = scaled_dataframe
else:
# ValueError: columns overlap but no suffix specified が発生するので、やむなく、'postgresql' を削除
del scaled_dataframe[primary_kw]
merged_df = merged_df.join(scaled_dataframe)
else:
print("---------------")
print("No Data: ", elems)
print("---------------")
if merged_df is not None:
results.append(merged_df.mean()/merged_df.mean().mean())
# データ整理
res_19 = pd.DataFrame(pd.DataFrame(results).sum())
res_19.columns = [ "trend"]
res_19["name"] = res_19.index
res_19.name = res_19.name.str.replace("'", "")
res_19 = res_19.reset_index(drop = True)
# 候補者データと検索トレンドをmerge
final_19 = pd.merge(data_19, res_19, on = "name", how = "left")
final_19= final_19.fillna(0)
final_19.head()
# 説明変数の作成
party_19 = pd.get_dummies(final_19["party"])
X_19 = pd.merge(final_19.drop(['name', 'party', 'district_J', 'num'],axis=1), party_19, left_index=True, right_index=True)
X_19.sex = (X_19.sex == "m")
X_19.head()
# 学習モデルを用いた当選確率の予測
prob = lr.predict_proba(X_19)[:, 1]
prob
array([0.158742 , 0.50235483, 0.3621762 , 0.20174061, 0.74296335,
0.70635213, 0.24673567, 0.37815342, 0.20505149, 0.06710085,
0.75969601, 0.59558288, 0.25242076, 0.75884715, 0.20369219,
0.20309021, 0.7613876 , 0.29612055, 0.82600095, 0.12008838,
0.20490012, 0.57170761, 0.24997479, 0.20429549, 0.52593273,
0.08597857, 0.23601316, 0.28800128, 0.57050237, 0.60183129,
0.12614082, 0.76884679, 0.20565778, 0.18877499, 0.04820443,
0.78260742, 0.29476714, 0.06745056, 0.37684378, 0.33056393,
0.85009167, 0.55467489, 0.20203994, 0.67937401, 0.81977391,
0.15074872, 0.06827313, 0.79655594, 0.68705021, 0.78649527,
0.20084461, 0.12580124, 0.25032349, 0.64301333, 0.49173961,
0.85346702, 0.91175473, 0.21155181, 0.33931208, 0.68340697,
0.16665054, 0.3751004 , 0.32879559, 0.20736463, 0.15087743,
0.61411738, 0.06238836, 0.20024893, 0.20278972, 0.31884302,
0.21926558, 0.19995159, 0.25007928, 0.75680284, 0.75390698,
0.14733936, 0.04900315, 0.31904478, 0.32045901, 0.20129224,
0.44880564, 0.20520294, 0.20339103, 0.87084304, 0.14059122,
0.65884537, 0.20550608, 0.09226068, 0.81332624, 0.20520294,
0.59625387, 0.75867713, 0.75884715, 0.59625387, 0.33221036,
0.20474884, 0.76189348, 0.82304443, 0.08715369, 0.20641751,
0.74784096, 0.20129224, 0.72717291, 0.20520294, 0.82331483,
0.59848791, 0.20459764, 0.75969601, 0.67998069, 0.73661433,
0.1230194 , 0.20429549, 0.81359819, 0.20189024, 0.73535093,
0.12382332, 0.20189024, 0.37778191, 0.32472147, 0.25743097,
0.81825942, 0.37662568, 0.20444653, 0.06803716, 0.53554602,
0.06721723, 0.28065732, 0.20490012, 0.08510667, 0.76273497,
0.20776107, 0.80207543, 0.16887818, 0.34771381, 0.20596142,
0.20369219, 0.55846774, 0.61784396, 0.06745056, 0.29592699,
0.68078864, 0.8905466 , 0.25921054, 0.20444653, 0.15182207,
0.397249 , 0.20144161, 0.20459764, 0.81218544, 0.29670169,
0.7770252 , 0.61937797, 0.33179837, 0.05489366, 0.79883739,
0.08590561, 0.25605973, 0.79962557, 0.8256001 , 0.08481782,
0.2065697 , 0.78870245, 0.65087186, 0.20490012, 0.4877603 ,
0.20174061, 0.34111322, 0.33336919, 0.20129224, 0.35584739,
0.80135571, 0.83145638, 0.32289109, 0.05065497, 0.59915733,
0.20189024, 0.76239862, 0.3286032 , 0.32167385, 0.08678491,
0.76020435, 0.20474884, 0.75178335, 0.13681086, 0.20490012,
0.37444747, 0.20414454, 0.12392412, 0.20369219, 0.81033657,
0.06571879, 0.65925995, 0.15039232, 0.81922433, 0.80676984,
0.65863372, 0.82452759, 0.29708948, 0.20520294, 0.7834533 ,
0.1090507 , 0.20596142, 0.79471198, 0.28840851, 0.20550608,
0.78434795, 0.18528275, 0.60004932, 0.32126863, 0.0873016 ,
0.75663055, 0.80307113, 0.32248508, 0.20084461, 0.32350062])