def __init__(self, cols=["x_resid", "y_resid", "phi_resid"],
min_num_obs=20,
separate_experiments=True,
separate_panels=True,
alpha=0.5,
max_n_groups=5,
min_group_size=300,
n_trials=500,
k1=2,
k2=2,
k3=100,
threshold_probability=0.975):
CentroidOutlier.__init__(self,
cols=cols,
min_num_obs=min_num_obs,
separate_experiments=separate_experiments,
separate_panels=separate_panels)
# Keep the FastMCD options here
self._alpha = alpha
self._max_n_groups = max_n_groups
self._min_group_size = min_group_size
self._n_trials = n_trials
self._k1 = k1
self._k2 = k2
self._k3 = k3
# Calculate Mahalanobis distance threshold
df = len(cols)
self._mahasq_cutoff = qchisq(threshold_probability, df)
return
def __init__(
self,
cols=None,
min_num_obs=20,
separate_experiments=True,
separate_panels=True,
block_width=None,
alpha=0.5,
max_n_groups=5,
min_group_size=300,
n_trials=500,
k1=2,
k2=2,
k3=100,
threshold_probability=0.975,
):
if cols is None:
cols = ["x_resid", "y_resid", "phi_resid"]
CentroidOutlier.__init__(
self,
cols=cols,
min_num_obs=min_num_obs,
separate_experiments=separate_experiments,
separate_panels=separate_panels,
block_width=block_width,
)
# Keep the FastMCD options here
self._alpha = alpha
self._max_n_groups = max_n_groups
self._min_group_size = min_group_size
self._n_trials = n_trials
self._k1 = k1
self._k2 = k2
self._k3 = k3
# Calculate Mahalanobis distance threshold
df = len(cols)
self._mahasq_cutoff = qchisq(threshold_probability, df)
return