-
Notifications
You must be signed in to change notification settings - Fork 0
/
inverse_problem.py
183 lines (123 loc) · 4.89 KB
/
inverse_problem.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
from file_utils import load_recommendations
def get_popularity_bias_denominator():
popularity_bias_denominator = 3
return popularity_bias_denominator
def get_popularity_bias_range():
popularity_bias_range = [-1, 0, 1, 2, 3]
return popularity_bias_range
def get_release_recency_bias_range():
release_recency_bias_range = [6, 12, 24, 36, 60, 120]
return release_recency_bias_range
def aggregate_recommendations(recommendations=None, verbose=False):
if recommendations is None:
recommendations = load_recommendations()
aggregated_recommendations = {}
for ranking in recommendations:
settings = ranking['algorithm_options']
popularity_bias = int(
get_popularity_bias_denominator() * float(settings['popularity_bias']),
)
release_recency_bias = int(settings['release_recency_bias'])
score_scale = float(ranking['score_scale'])
for app_id, score in zip(ranking['app_ids'], ranking['scores']):
tweaked_output = score_scale * score
current_data = {
"popularity_bias": popularity_bias,
"release_bias": release_recency_bias,
"tweaked_score": tweaked_output,
}
try:
aggregated_recommendations[str(app_id)].append(current_data)
except KeyError:
aggregated_recommendations[str(app_id)] = [current_data]
if verbose:
print(f'#recommended apIDs = {len(aggregated_recommendations)}')
return aggregated_recommendations
def count_rankings(recommendations=None, verbose=False):
if recommendations is None:
recommendations = load_recommendations()
first_ranking_index = 0
num_rankings = len(recommendations)
ranking_size = len(recommendations[first_ranking_index]['app_ids'])
if verbose:
print(f'There are {num_rankings} rankings of {ranking_size} apps.')
return num_rankings, ranking_size
def count_occurrences(aggregated_recommendations, verbose=False):
stats = {}
for app_id, occurrences in aggregated_recommendations.items():
num_occurrences = len(occurrences)
try:
stats[num_occurrences].append(app_id)
except KeyError:
stats[num_occurrences] = [app_id]
if verbose:
print('How many apps appear in n rankings?')
for i in sorted(stats.keys()):
num_apps = len(stats[i])
print(f'[n = {i:2} occurrences] {num_apps:3} apps.')
total_num_apps = get_total_num_apps(stats, verbose=verbose)
total_num_occurrences = get_total_num_occurrences(stats, verbose=verbose)
return stats
def get_total_num_apps(stats, verbose=False):
total_num_apps = sum(len(l) for l in stats.values())
if verbose:
print(f'Total: {total_num_apps:5} apps.')
return total_num_apps
def get_total_num_occurrences(stats, verbose=False):
total_num_occurrences = sum(n * len(l) for (n, l) in stats.items())
if verbose:
print(f'Total: {total_num_occurrences:5} occurrences.')
return total_num_occurrences
def summarize_occurrences(
aggregated_recommendations,
stats=None,
chosen_num_occurrences=None,
verbose=False,
):
if stats is None:
stats = count_occurrences(aggregated_recommendations, verbose=verbose)
if chosen_num_occurrences is None:
chosen_num_occurrences = max(stats.keys())
app_ids = stats[chosen_num_occurrences]
pb_val = get_popularity_bias_range()
rb_val = get_release_recency_bias_range()
pb_occurrences_dict = {}
rb_occurrences_dict = {}
for app_id in app_ids:
val = aggregated_recommendations[app_id]
pb_occurrences = [0] * len(pb_val)
rb_occurrences = [0] * len(rb_val)
for elem in val:
pb = elem['popularity_bias']
rb = elem['release_bias']
pb_occurrences[pb_val.index(pb)] += 1
rb_occurrences[rb_val.index(rb)] += 1
pb_occurrences_dict[app_id] = pb_occurrences
rb_occurrences_dict[app_id] = rb_occurrences
if verbose:
print(f'AppIDs = {app_ids}')
print(
'popularity bias = {} ; #occurrences = {}'.format(
pb_val,
pb_occurrences_dict,
),
)
print(
'release recency bias = {} ; #occurrences = {}'.format(
rb_val,
rb_occurrences_dict,
),
)
return app_ids, pb_occurrences_dict, rb_occurrences_dict
def main():
aggregated_recommendations = aggregate_recommendations(verbose=True)
stats = count_occurrences(aggregated_recommendations, verbose=True)
app_ids, pb_occurrences_dict, rb_occurrences_dict = summarize_occurrences(
aggregated_recommendations,
stats,
chosen_num_occurrences=max(stats.keys()),
verbose=True,
)
return
if __name__ == '__main__':
main()