def prepare_data(clinical_data,
display_cmat=True,
VIF=10,
skip_variable_reduction=False):
#VIF of 10 means that all the features with R^2 above 0.9 will removed
time, event, fdata = extract_time_event_and_features(clinical_data)
if not skip_variable_reduction:
fdata2 = calculate_vif_(fdata, VIF)
else:
fdata2 = fdata
fdata2 = remove_nan_correlations(fdata2, time)
if display_cmat:
from pysurvival.utils.display import correlation_matrix
correlation_matrix(fdata2, figure_size=(30, 15), text_fontsize=10)
return fdata2, time, event
cph = CoxPHFitter()
data = train[albumin_change]
cph.fit(data, 'PFS', event_col='disease_progress')
cph.print_summary()
cph = CoxPHFitter()
data = train[haemoglobin_change]
cph.fit(data, 'PFS', event_col='disease_progress')
cph.print_summary()
# In[ ]:
# In[34]:
correlation_matrix(df[features], figure_size=(30, 15), text_fontsize=10)
# In[ ]:
# In[35]:
def compute_scores(model, table, timepoints, variables):
c_indexes = []
for i in timepoints:
table.loc[:, 'disease_progress_temp'] = table['disease_progress']
table.loc[:, 'PFS_temp'] = table['PFS']
table.loc[table.PFS > i, 'disease_progress_temp'] = 0
table.loc[table.PFS > i, 'PFS_temp'] = i
c_indexes.append(
concordance_index(model, table[variables], table['PFS_temp'],
# Removing duplicates if there exist
N_dupli = sum(df.duplicated(keep='first'))
df = df.drop_duplicates(keep='first').reset_index(drop=True)
print("The raw_dataset contains {} duplicates".format(N_dupli))
# Number of samples in the dataset
N = df.shape[0]
df.columns
from pysurvival.utils.display import correlation_matrix
correlation_matrix(df[features], figure_size=(10,10), text_fontsize=8)
to_remove = ['totalJogos', 'idaEstadio']
features = np.setdiff1d(features, to_remove).tolist()
# Building training and testing sets
from sklearn.model_selection import train_test_split
index_train, index_test = train_test_split( range(N), test_size = 0.4)
data_train = df.loc[index_train].reset_index( drop = True )
data_test = df.loc[index_test].reset_index( drop = True )
# Creating the X, T and E inputs
X_train, X_test = df[features], data_test[features]
T_train, T_test = df[time_column], data_test[time_column]