-
Notifications
You must be signed in to change notification settings - Fork 0
/
score_model.py
148 lines (115 loc) · 3.96 KB
/
score_model.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
import numpy as np
import pandas as pd
import os
import time
import sys
import pickle
MODEL_FILE = 'RF_model_trained.sav'
MODEL = pickle.load(open(MODEL_FILE, 'rb'))
SCORE_MAX_NO_SYMPTOMS = 50
SCORE_WORST_CASE = 10
work_env_mapping = { 'healthcare' : 0, 'close_contact' : 1, 'regular_contact' : 2, 'no_contact' : 3 }
worst_case = pd.Series( {
'generally_ill' : 1, # Q1
'resp_cough' : 1, # Q2.1
'resp_throat' : 1, # Q2.2
'resp_breath' : 1, # Q2.3
'taste_loss' : 1, # Q3
'sympt_fever' : 1, # Q4.1
'sympt_sens_fev': 1, # Q4.2
'sympt_musc' : 1, # Q4.3
'sympt_head' : 1, # Q4.4
'work_env' : 3, # Q6
'proxim_sympt' : 1, # Q8.1
'proxim_case' : 1, # Q8.2
'contact_sympt' : 1, # Q9.1
'contact_case' : 1, # Q9.2
} )
no_symptoms = pd.Series( {
'resp_cough' : 1, # Q2.1
'resp_throat' : 1, # Q2.2
'resp_breath' : 1, # Q2.3
'taste_loss' : 1, # Q3
'sympt_fever' : 1, # Q4.1
'sympt_sens_fev': 1, # Q4.2
'sympt_musc' : 1, # Q4.3
'sympt_head' : 1, # Q4.4
} )
def validate_data(sample):
# adjust hidden answers to defaults:
if sample.generally_ill == 1:
sample['illness_level'] = -1
if sample.work_env == 3:
sample['work_protec'] = -1
sample['work_exposure'] = -1
return sample
def convert_input(args, result_dict):
for key, arg in args.items():
if arg == 'yes':
result_dict[key] = 0
elif arg == 'no':
result_dict[key] = 1
elif arg == 'unknown':
result_dict[key] = -1
elif key == 'work_env':
result_dict[key] = work_env_mapping[arg]
elif key == 'illness_level':
result_dict[key] = int(arg)
else:
print('DATA UNKNOWN:', key, arg)
print(key, arg)
return validate_data( pd.Series(result_dict) )
def initialize_dict():
return {
'generally_ill' : 1, # Q1
'illness_level' : 0, # NEW
'resp_cough' : 1, # Q2.1
'resp_throat' : 1, # Q2.2
'resp_breath' : 1, # Q2.3
'taste_loss' : 1, # Q3
'sympt_fever' : 1, # Q4.1
'sympt_sens_fev': 1, # Q4.2
'sympt_musc' : 1, # Q4.3
'sympt_head' : 1, # Q4.4
'date_symptoms' : 0, # Q5
'work_env' : 3, # Q6
'work_protec' : 1, # Q7.a
'work_exposure' : 1, # Q7.b
'proxim_sympt' : 1, # Q8.1
'proxim_case' : 1, # Q8.2
'contact_sympt' : 1, # Q9.1
'contact_case' : 1, # Q9.2
}
def predict(in_data):
# Function for predicting one sample at a time (passed as pandas series)
features = in_data.values.reshape(1, -1)[:,2:]
print('Predicting for: ', features)
prediction = MODEL.predict(features)
pred_proba = MODEL.predict_proba(features)
return prediction[0], pred_proba[0]
def score_to_percentage(score, probabilities = None):
p0 = [20, 40, 60, 80]
p_cls = [35, 50, 55, 60]
if probabilities is None:
return score * 20
cls_idx = int(score) - 1
highest_proba = probabilities[cls_idx] * 100
cls_2nd = np.argsort(probabilities)[-2]
if highest_proba >= p_cls[cls_idx]:
return p0[cls_idx]
else:
return (p0[cls_idx] + p0[cls_2nd]) / 2
def scorer(args):
print('ENTERING SCORER')
results = convert_input( args, initialize_dict() )
if (results[worst_case.index] == worst_case).all(): # hard-code the absolute worst case
print('Detected no symptoms; returning')
return int(SCORE_WORST_CASE / 10)
prediction, probabilities = predict(results)
score = score_to_percentage(prediction, probabilities)
if (results[no_symptoms.index] == no_symptoms).all():
print('Capped at %d' %SCORE_MAX_NO_SYMPTOMS)
score = min(score, SCORE_MAX_NO_SYMPTOMS)
print('LEAVING SCORER')
print('Predicted score: %d' %score)
return int(score/10)