rotor_clustering
DBSCANRotorClustering
Bases: RotorClustering
Clusters rotors with positive/negative rotation direction seperately in time and space simultaneously using the DBSCAN algorithm.
The t ime and space feature spaces are rescaled by a certain weight to yield more importance to one or the other.
coords = coords
instance-attribute
model = DBSCAN(eps=eps, min_samples=min_samples)
instance-attribute
space_scale = 1
instance-attribute
space_weight = 1
instance-attribute
time_scale = 1
instance-attribute
time_weight = 1
instance-attribute
__init__(rotors, coords, eps=2, min_samples=3)
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
eps |
float
|
Clustering radius used in DBSCAN. Default to 2. |
2
|
min_samples |
int
|
Min amount of cluster samples required for a point that is part if this cluster to be classified as a core point. Defaults to 3. |
3
|
cluster()
Clusters the rotors with positive/negative rotation direction separately in time and space simultaneously.
Returns:
| Type | Description |
|---|---|
DataFrame
|
DataFrame with columns ['t_ind', 'x', 'y', 'z', 'direction', 'cluster', 'cluster_size']. |
project_on_mesh(rotors_cl)
Projects rotor positions on mesh by replacing with the nearest mesh node.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
rotors_cl |
DataFrame
|
DataFrame with columns ['t_ind', 'x', 'y', 'z', 'direction', 'cluster', 'cluster_size']. |
required |
Returns:
| Type | Description |
|---|---|
DataFrame
|
pd.DataFrame: DataFrame with columns ['t_ind', 'node_ind', 'direction', 'cluster', 'cluster_size']. |
reduce_clusters(rotors_cl)
Reduces rotors to the cluster center for each time.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
rotors_cl |
DataFrame
|
Dataframe with columns ['t_ind', 'x', 'y', 'z', 'direction', 'cluster', 'cluster_size']. |
required |
Returns:
| Type | Description |
|---|---|
DataFrame
|
pd.DataFrame: Dataframe with columns ['t_ind', 'x', 'y', 'z', 'direction', 'cluster', 'cluster_size']. |
set_space_scale(space_scale)
Distance between nodes Default: 1
set_space_weight(space_weight)
Space feature space scaling factor Default: 1
set_time_scale(time_scale)
Time between frames Default: 1
set_time_weight(time_weight)
Time feature space weight factor Default: 1