def predict_surv_df(self,
input,
batch_size=8224,
eval_=True,
to_cpu=False,
num_workers=0):
"""Predict the survival function for `input` and return as a pandas DataFrame.
See `predict_surv` to return tensor or np.array instead.
Arguments:
input {tuple, np.ndarray, or torch.tensor} -- Input to net.
Keyword Arguments:
batch_size {int} -- Batch size (default: {8224})
eval_ {bool} -- If 'True', use 'eval' mode on net. (default: {True})
num_workers {int} -- Number of workers in created dataloader (default: {0})
Returns:
pd.DataFrame -- Predictions
"""
surv = self.predict_surv(input, batch_size, True, eval_, to_cpu,
num_workers)
index = None
if self.duration_index is not None:
index = utils.make_subgrid(self.duration_index, self.sub)
return pd.DataFrame(surv.transpose(), index)
def predict_surv_df(self,
input,
batch_size=8224,
eval_=True,
num_workers=0):
self._check_out_features()
surv = self.predict_surv(input, batch_size, True, eval_, True,
num_workers)
index = None
if self.duration_index is not None:
index = make_subgrid(self.duration_index, self.sub)
return pd.DataFrame(surv.transpose(), index)
def test_make_subgrid(sub, start, stop, n):
grid = np.linspace(start, stop, n)
new_grid = make_subgrid(grid, sub)
true_new = np.linspace(start, stop, n * sub - (sub - 1))
assert len(new_grid) == len(true_new)
assert np.abs(true_new - new_grid).max() < 1e-13
def test_make_subgrid_1(n):
grid = np.random.uniform(0, 100, n)
grid = np.sort(grid)
new_grid = make_subgrid(grid, 1)
assert len(new_grid) == len(grid)
assert (new_grid == grid).all()
