PranjalDureja0002

inpatient_data['Admission_Date'] = pd.to_datetime(inpatient_data['AdmissionDt'] , format = '%Y-%m-%d')
inpatient_data['Discharge_Date'] = pd.to_datetime(inpatient_data['DischargeDt'],format = '%Y-%m-%d')

inpatient_data['ClaimStart_Date'] = pd.to_datetime(inpatient_data['ClaimStartDt'] , format = '%Y-%m-%d')
inpatient_data['ClaimEnd_Date'] = pd.to_datetime(inpatient_data['ClaimEndDt'],format = '%Y-%m-%d')

inpatient_data['DOB'] = pd.to_datetime(inpatient_data['DOB'] , format = '%Y-%m-%d')
inpatient_data['DOD'] = pd.to_datetime(inpatient_data['DOD'],format = '%Y-%m-%d')
inpatient_data['Age'] = round(((inpatient_data['ClaimStart_Date'] - inpatient_data['DOB']).dt.days + 1)/365.25)

PranjalDureja0002

plt.figure(figsize=(15, 9))
sns.pointplot(x = outpatients_.physician_count, y = outpatients_.id_Count,hue=outpatients_.PotentialFraud)
plt.suptitle('Physician_attended vs Beneficiaries_count\n')

sns.pointplot(x = inpatients_.physician_count, y = inpatients_.id_Count,hue=inpatients_.PotentialFraud)
plt.suptitle('Physician_attended vs Beneficiaries_count\n')
plt.show()

PranjalDureja0002

physician_count = outpatients_['AttendingPhysician'].value_counts().to_dict()
outpatients_['physician_count']=outpatients_['AttendingPhysician'].map(physician_count)

physician_count = inpatients_['AttendingPhysician'].value_counts().to_dict()
inpatients_['physician_count']=inpatients_['AttendingPhysician'].map(physician_count)

ax = sns.countplot(y='AttendingPhysician',data=outpatients_,hue='PotentialFraud',order = outpatients_['AttendingPhysician'].value_counts().head(20).index)
d1 = outpatients_['AttendingPhysician'].value_counts().to_dict()
s_s1 = sum(list(d1.values()))
for p in ax.patches:

PranjalDureja0002

colors_list = ['#5bc0de','#d9534f']
plt.title("Potential Fraud distribution")
ax = patient_data['PotentialFraud'].value_counts().plot(kind='bar', figsize=(5,5), width=0.8,color = colors_list)
total = len(patient_data['PotentialFraud'])
for p in ax.patches:
        percentage = '{:.1f}%'.format(100 * p.get_height()/total)        
        x = p.get_x() + p.get_width() - 0.5
        y = p.get_y() + p.get_height()
        ax.annotate(percentage, (x, y))
plt.xlabel('Potential_Fraud')

PranjalDureja0002

BeneID                               0.000000
ClaimID                              0.000000
ClaimStartDt                         0.000000
ClaimEndDt                           0.000000
Provider                             0.000000
InscClaimAmtReimbursed               0.000000
AttendingPhysician                   0.270149
OperatingPhysician                  79.497538
OtherPhysician                      64.218548

PranjalDureja0002

per_miss_values = patient_data.isnull().sum() * 100 / len(patient_data)

PranjalDureja0002

patient_data_io = pd.merge(train_d_inpatient, train_d_outpatient, left_on = [ col for col in train_d_outpatient.columns if col in train_d_inpatient.columns], \
                    right_on = [ col for col in train_d_outpatient.columns if col in train_d_inpatient.columns], how = 'outer')

patient_data = pd.merge(patient_data_io,train_d_beneficiary,how='inner',on='BeneID' ).\
              merge(label_d_data,how='outer',on='Provider')

PranjalDureja0002

	providerID	y_predicted
0	PRV57070	  0
1	PRV57070	  1
2	PRV57070	  0
3	PRV57070	  0
4	PRV57070	  0
5	PRV57070	  1
6	PRV57070	  0
7	PRV57070	  1
8	PRV57070	  0

PranjalDureja0002

def final_fun_1(X):
    """ function takes raw data as input,preprocessing is done,
    feature engineering is performed and predictions made on the
    best model already trained"""
            
    d_beneficiary = pd.read_csv('health_cs_data/' + X[0])
    d_inpatient = pd.read_csv('health_cs_data/' + X[1])
    d_outpatient = pd.read_csv('health_cs_data/' + X[2])
    d_labels = pd.read_csv('health_cs_data/' + X[3])
    

PranjalDureja0002

def find_best_threshold(threshold, fpr, tpr):
    t = threshold[np.argmax(tpr*(1-fpr))]     
    print("the maximum value of tpr*(1-fpr)", max(tpr*(1-fpr)), "for threshold", np.round(t,3))
    return t

def predict_with_best_t(proba, threshold):
    predictions = []
    for i in proba:                              
        if i>=threshold:
            predictions.append(1)