-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathMIMIC.py
132 lines (98 loc) · 3.92 KB
/
MIMIC.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
import pandas as pd
import numpy as np
import os
from gensim.models import Word2Vec
from gensim.models.word2vec import LineSentence
import os
import numpy as np
import pandas as pd
import tensorflow as tf
from tensorflow.contrib import learn
from sklearn.metrics import *
from tensorflow.contrib.layers import l2_regularizer
ICD_path = '/Users/liangyunfan/Downloads/MIMIC3_data/D_ICD_DIAGNOSES.csv'
diagnose_path = '/Users/liangyunfan/Desktop/DIAGNOSES_ICD_filter.csv'
prescription_path = '/Users/liangyunfan/Downloads/PRESCRIPTIONS_1.csv'
ICD_CODE = pd.read_csv(diagnose_path)
ICD_CODE = ICD_CODE['ICD9_CODE']
ICD_CODE = set([i for i in ICD_CODE])
ICD_CODE = np.array([i for i in ICD_CODE])
ICD_CODE_dict = {}
for i,j in enumerate(ICD_CODE):
ICD_CODE_dict[j] = i
import pickle as plk
with open('ICDCODEDICR.plk','wb') as f:
plk.dump(ICD_CODE_dict,f)
ICD_num = 6985
def gen_onehot():
diagnose_data = pd.read_csv(diagnose_path)
prescription_data = pd.read_csv(prescription_path)
HADM_ID_p = prescription_data['HADM_ID']
HADM_ID_p = [i for i in HADM_ID_p]
print(np.array([i for i in set(HADM_ID_p)]).shape)
Medicine = prescription_data['DRUG']
Medicine_set = [i for i in set(Medicine)]
Med_dict = {}
for i,j in enumerate(Medicine_set):
Med_dict[j] = i
Med_num = np. array(Medicine_set).shape[0]
print('number of medicine is: {}'.format(Med_num))
Med_HA_dict = {}
hold = 0
for i in range(np.array(HADM_ID_p).shape[0]):
if hold == 0 :
add = np.zeros([1,Med_num])
add[0,Med_dict.get(Medicine[i])] = 1
if HADM_ID_p[i] == hold:
add[0,Med_dict.get(Medicine[i])] = 1
if HADM_ID_p[i] != hold and hold != 0:
Med_HA_dict[HADM_ID_p[i - 1]] = add
add = np.zeros([1, Med_num])
add[0, Med_dict.get(Medicine[i])] = 1
hold = HADM_ID_p[i]
print(np.array([i for i in Med_HA_dict.keys()]).shape)
# print(Med_HA_dict)
# diagnose_data.sort_values('SUBJECT_ID',inplace=True,ascending=True)
SUBJECT_ID = diagnose_data['SUBJECT_ID']
SUBJECT_ICD = diagnose_data['ICD9_CODE']
HADM_ID = diagnose_data['HADM_ID']
HADM_ID = [int(i)for i in HADM_ID]
HADM_ID_set = [i for i in set(HADM_ID)]
print(np.array(HADM_ID_set).shape)
PATIENT_NUM = np.array([i for i in set(SUBJECT_ID)]).shape[0]
print('the number of patient is: {}'.format(PATIENT_NUM))
print('the ICD occured is: {}'.format(ICD_num))
DIAGNOSE = np.zeros([1, ICD_num + Med_num])
hold = 0
for x in range(SUBJECT_ID.shape[0]):
if HADM_ID[x] in Med_HA_dict.keys():
if hold != HADM_ID[x] and hold == 0:
add = np.zeros([1,ICD_num])
add[0,ICD_CODE_dict.get(SUBJECT_ICD[x])] = 1
if hold != HADM_ID[x] and hold != 0:
add = np.concatenate((add, Med_HA_dict.get(HADM_ID[x])), axis=1)
DIAGNOSE = np.concatenate((DIAGNOSE, add), axis=0)
add = np.zeros([1,ICD_num])
add[0,ICD_CODE_dict.get(SUBJECT_ICD[x])] = 1
if hold == HADM_ID[x]:
add[0,ICD_CODE_dict.get(SUBJECT_ICD[x])] = 1
hold = HADM_ID[x]
# if x == 200000:
# break
if x%1000==0:
print(x,'\n')
DIAGNOSE = DIAGNOSE[1:]
print('one hot data is:{}'.format(DIAGNOSE.shape))
print('test: ',DIAGNOSE.sum(axis = 1))
np.save('DIAGNOSE.npy',DIAGNOSE)
def Word2Vec_train():
ICD_word_vec = [x for x in ICD_CODE]
ICD_word_vec = [y for y in map(lambda x:str(x), ICD_word_vec)]
ICD_word_vec = [z+'\n' for z in ICD_word_vec]
with open('MIMIC_ICD.txt','w')as f:
f.writelines(ICD_word_vec)
model = Word2Vec(LineSentence('MIMIC_ICD.txt'), size=512, window=5, min_count=1, workers=4, iter=10000)
# model.save("MIMIC_ICD_512.model")
model.wv.save_word2vec_format("MIMIC_Word2Vec.vector",binary=False)
# gen_onehot()
# Word2Vec_train()