Constants

survivors.constants.RANDOM_STATE = 123

int: Fixed seed for model reproducibility

survivors.constants.TIME_NAME = 'time'

str: Fixed names for internal representation

survivors.constants.get_bins(time, cens=None, mode='a', num_bins=100)

Generate array of time points in timeline (from sample)

Parameters

timearray-like

Time of occurred events.

censarray-like, optional

Censoring flags. The default is None (all events occurred).

modestr, optional

Method of generation. The default is ‘a’. ‘a’ : all points (from min to max) ‘q’ : quantile points (quantity is based on num_bins)

num_binsint, optional

Quantity of required points. The default is 100.

Returns

binsarray

Timeline

survivors.constants.get_y(cens, time)

Internal representation of target variable

Parameters

censarray-like, shape = (n_events,)

Censoring flags.

timearray-like, shape = (n_events,)

Time of occurred events.

Returns

ystructured array

Output containing the binary event indicator as first field, and time of event or time of censoring as second field.

survivors.constants.pd_to_xy(df)

Splitting pandas dataframe to feature space and target variables

Parameters

dfpandas DataFrame

Must contain columns CENS_NAME and TIME_NAME.

Returns

Xpandas DataFrame

Feature space with remaining features.

ystructured array

y containing the binary event indicator as first field, and time of event or time of censoring as second field.