VotePredictor/baseline.py at master · mstfbl/VotePredictor · GitHub

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
206
207
208
209
import os
import json
import nltk
import time
import random
import matplotlib.pyplot

def main():

  with open('training_set.json') as d:
    training_set = json.load(d)

  with open('validation_set.json') as d:
    validation_set = json.load(d)

  # Weighted flip baseline
  model = train_weighted_flip(training_set)
  results = validate_weighted_flip(validation_set, model)
  legislator_success = list()
  for legislator in results[0]:
    legislator_success.append(results[0][legislator].get("success", 0) / float(results[0][legislator]["total"]))

  print(sum(legislator_success)/ len(legislator_success))
  correct = 0
  for vote in results[1]:
    if results[1][vote]:
      correct += 1
  print ("Correct: " + str(float(correct) / len(results[1])))
  with open('baseline_weightedflip.txt', 'w') as f:
    for p in legislator_success:
      f.write("%s\n" % str(p))
    f.write("%s\n" % ("Legislator Avg: " + str(sum(legislator_success)/ len(legislator_success))))
    f.write("%s\n" % ("Bills Correct: " + str(float(correct) / len(results[1]))))

  # All Yea Baseline
  results = validate_yea(validation_set)
  legislator_success = list()
  for legislator in results[0]:
    legislator_success.append(results[0][legislator].get("success", 0) / float(results[0][legislator]["total"]))

  print(sum(legislator_success)/ len(legislator_success))
  correct = 0
  for vote in results[1]:
    if results[1][vote]:
      correct += 1
  print ("Correct: " + str(float(correct) / len(results[1])))
  with open('baseline_yea.txt', 'w') as f:
    for p in legislator_success:
      f.write("%s\n" % str(p))
    f.write("%s\n" % ("Legislator Avg: " + str(sum(legislator_success)/ len(legislator_success))))
    f.write("%s\n" % ("Bills Correct: " + str(float(correct) / len(results[1]))))

  # All Nay Baseline
  results = validate_nay(validation_set)
  legislator_success = list()
  for legislator in results[0]:
    legislator_success.append(results[0][legislator].get("success", 0) / float(results[0][legislator]["total"]))

  print(sum(legislator_success)/ len(legislator_success))
  correct = 0
  for vote in results[1]:
    if results[1][vote]:
      correct += 1
  print ("Correct: " + str(float(correct) / len(results[1])))
  with open('baseline_nay.txt', 'w') as f:
    for p in legislator_success:
      f.write("%s\n" % str(p))
    f.write("%s\n" % ("Legislator Avg: " + str(sum(legislator_success)/ len(legislator_success))))
    f.write("%s\n" % ("Bills Correct: " + str(float(correct) / len(results[1]))))


def train_weighted_flip(training_set):
  model = {}
  for vote in training_set:

    def trainer(givenVote):
        if givenVote in training_set[vote]["votes"]:
          for legislator in training_set[vote]["votes"][givenVote]:
            # group Aye with Yea and No with Nay
            loggedCategory = givenVote
            if givenVote == "Aye":
              loggedCategory = "Yea"
            elif givenVote == "No":
              loggedCategory = "Nay"

            # initialize new congressman
            if legislator["id"] not in model:
              model[legislator["id"]] = {}

            # increment word count for a given label
            model[legislator["id"]][loggedCategory] = model[legislator["id"]].get(loggedCategory, 0) + 1

    trainer("Nay")
    trainer("No")
    trainer("Yea")
    trainer("Aye")
    trainer("Not Voting")
  return model

def validate_weighted_flip(validation_set, model):
  vote_results = {}
  legislator_results = {}
  for vote in validation_set:
    vote_count = [0,0,0]

    def validateXVotes(givenVote,givenLabel):
      if givenVote in validation_set[vote]["votes"]:
        for legislator in validation_set[vote]["votes"][givenVote]:
          if legislator["id"] not in model:
            dprint("Congressman not seen in training set preset in model test - " + givenVote)
            continue
          label = generate_label(model[legislator["id"]], vote_count)
          if legislator["id"] not in legislator_results:
            legislator_results[legislator["id"]] = {}
          #If predicted correctly
          if label == givenLabel:
            legislator_results[legislator["id"]]["success"] = legislator_results[legislator["id"]].get("success", 0) + 1
          legislator_results[legislator["id"]]["total"] = legislator_results[legislator["id"]].get("total", 0) + 1

    #Validating votes predicted
    validateXVotes("Nay",0)
    validateXVotes("No",0)
    validateXVotes("Yea",1)
    validateXVotes("Aye",1)
    validateXVotes("Not Voting",2)

    model_result = (vote_count[1] / float(vote_count[0] + vote_count[1])) >= validation_set[vote]["requires"]
    if model_result == validation_set[vote]["result"]:
      vote_results[vote] = True
    else:
      vote_results[vote] = False

  return [legislator_results, vote_results]

def validate_yea(validation_set):
  vote_results = {}
  legislator_results = {}
  for vote in validation_set:
    def validateXVotes(givenVote,givenLabel):
      if givenVote in validation_set[vote]["votes"]:
        for legislator in validation_set[vote]["votes"][givenVote]:
          if legislator["id"] not in legislator_results:
            legislator_results[legislator["id"]] = {}
          #If predicted correctly
          if givenVote == "Yea" or givenVote == "Aye":
            legislator_results[legislator["id"]]["success"] = legislator_results[legislator["id"]].get("success", 0) + 1
          legislator_results[legislator["id"]]["total"] = legislator_results[legislator["id"]].get("total", 0) + 1

    #Validating votes predicted
    validateXVotes("Nay",0)
    validateXVotes("No",0)
    validateXVotes("Yea",1)
    validateXVotes("Aye",1)
    validateXVotes("Not Voting",2)

    model_result = True
    if model_result == validation_set[vote]["result"]:
      vote_results[vote] = True
    else:
      vote_results[vote] = False

  return [legislator_results, vote_results]

def validate_nay(validation_set):
  vote_results = {}
  legislator_results = {}
  for vote in validation_set:
    def validateXVotes(givenVote,givenLabel):
      if givenVote in validation_set[vote]["votes"]:
        for legislator in validation_set[vote]["votes"][givenVote]:
          if legislator["id"] not in legislator_results:
            legislator_results[legislator["id"]] = {}
          #If predicted correctly
          if givenVote == "Nay" or givenVote == "No":
            legislator_results[legislator["id"]]["success"] = legislator_results[legislator["id"]].get("success", 0) + 1
          legislator_results[legislator["id"]]["total"] = legislator_results[legislator["id"]].get("total", 0) + 1

    #Validating votes predicted
    validateXVotes("Nay",0)
    validateXVotes("No",0)
    validateXVotes("Yea",1)
    validateXVotes("Aye",1)
    validateXVotes("Not Voting",2)

    model_result = True
    if model_result == validation_set[vote]["result"]:
      vote_results[vote] = True
    else:
      vote_results[vote] = False

  return [legislator_results, vote_results]

def generate_label(legislator, vote_count):
  p_nay = legislator.get("Nay", 0) / float(legislator.get("Nay", 0) + legislator.get("Yea", 0)  + legislator.get("Not Voting", 0))
  p_yea = legislator.get("Yea", 0) / float(legislator.get("Nay", 0) + legislator.get("Yea", 0) + legislator.get("Not Voting", 0))
  r = random.random()

  if r < p_nay:
    vote_count[0] += 1
    return 0
  elif r < p_nay + p_yea:
    vote_count[1] += 1
    return 1
  else:
    vote_count[2] += 1
    return 2

if __name__ == "__main__":
  main()