def get_linkage(self, metric="sspd", method="ward", type_d="euclidean"):
traj_list = []
for traj in self.lat.columns:
ds = pd.DataFrame([])
ds["lat"] = self.lat[traj]
ds["lon"] = self.lon[traj]
traj_list.append(ds.values)
self.p_dist = tdist.pdist(traj_list, metric=metric, type_d=type_d)
self.link = fc.linkage(self.p_dist, method=method)
return self
def get_space_sim(d):
coors = [rdp(t[xy_cols].values, 1e-4) for t in d]
spatial_mat = np.identity(n)
spatial_dist = dist.pdist(coors, dist_measure)
min = spatial_dist.min()
spatial_dist = np.divide(spatial_dist - min, spatial_dist.max() - min)
k = 0
for i in range(n):
for j in range(i + 1, n):
spatial_mat[i, j] = spatial_mat[j, i] = spatial_dist[k]
k += 1
# print(spatial_dist)
return np.divide(1, 1 + spatial_mat)
import pandas as pd
import numpy as np
traj_list = pickle.load(
open(
"/Users/bguillouet/These/trajectory_distance/data/benchmark_trajectories.pkl",
"rb"))[:100]
time_dict = collections.defaultdict(dict)
for distance in [
"sspd", "frechet", "discret_frechet", "hausdorff", "dtw", "lcss",
"edr", "erp"
]:
t_euclidean = timeit.timeit(
lambda: tdist.pdist(traj_list, metric=distance), number=1)
if not (distance in ["frechet", "discret_frechet"]):
t_spherical = timeit.timeit(lambda: tdist.pdist(
traj_list, metric=distance, type_d="spherical"),
number=1)
else:
t_spherical = -1
time_dict[distance] = {"Euclidean": t_euclidean, "Spherical": t_spherical}
t_cells_conversion_dic = collections.defaultdict(int)
for precision in [5, 6, 7]:
cells_list_, _, _, _, _ = trajectory_set_grid(traj_list,
precision=precision)
cells_list = map(lambda x: np.array(x)[:, :2], cells_list_)
print('--', i)
name = namelist[i]
if (label[name] == 1).sum() > 0:
trajs = resample(data[name])
trajs_norm = []
for t in trajs:
t[:, 0] /= 960.
t[:, 1] /= 540.
trajs_norm.append(t)
trajs = np.array(trajs_norm)
traj_num = len(trajs)
y = sd.pdist(
trajs.transpose(0, 2, 1).reshape(traj_num, -1), 'euclidean')
dm_euc = sd.squareform(y)
y = tdist.pdist(trajs, metric='dtw')
dm_dtw = sd.squareform(y)
y = tdist.pdist(trajs, metric='sspd')
dm_sspd = sd.squareform(y)
y = tdist.pdist(trajs, metric='lcss', eps=0.05)
dm_lcss = sd.squareform(y)
y = tdist.pdist(trajs, metric='edr', eps=0.05)
dm_edr = sd.squareform(y)
y = tdist.pdist(trajs, metric='erp', g=np.zeros(2, dtype=float))
dm_erp = sd.squareform(y)
y = tdist.pdist(trajs, metric='frechet')
dm_fre = sd.squareform(y)
y = tdist.pdist(trajs, metric='hausdorff')
dm_hau = sd.squareform(y)
dms_euc.append(dm_euc)
import numpy as np
import traj_dist.distance as tdist
import pickle
traj_list = pickle.load(open("/Users/bguillouet/These/trajectory_distance/data/benchmark_trajectories.pkl", "rb"))[:10]
traj_A = traj_list[0]
traj_B = traj_list[1]
# Simple distance
dist = tdist.sspd(traj_A, traj_B)
print(dist)
# Pairwise distance
pdist = tdist.pdist(traj_list, metric="sspd")
print(pdist)
# Distance between two list of trajectories
cdist = tdist.cdist(traj_list, traj_list, metric="sspd")
print(cdist)
def __initial_match(self,
candidate_list: (np.ndarray, np.generic),
min_pts=2,
t=50,
criterion='distance'):
# TODO group matching for non-grouped user
# 1 : dbscan algorithm + gps based movement vector alignment -> clear!
# 2 : acceleration -> let's discuss
"""Performs initial-clustering on cn candidate_list(nT x 2 numpy array) and returns group lists.
Parameters
----------
candidate_list : array of shape (n_samples, n_of_time_steps, pair of latitude and longitude
min_pts : minimum members of a group for HDBSCAN-algorithm
t : scalar
For criteria 'inconsistent', 'distance' or 'monocrit',
this is the threshold to apply when forming flat clusters.
For 'maxclust' or 'maxclust_monocrit' criteria,
this would be max number of clusters requested.
criterion : str, optional
The criterion to use in forming flat clusters. This can
be any of the following values:
``inconsistent`` :
If a cluster node and all its
descendants have an inconsistent value less than or equal
to `t` then all its leaf descendants belong to the
same flat cluster. When no non-singleton cluster meets
this criterion, every node is assigned to its own
cluster. (Default)
``distance`` :
Forms flat clusters so that the original
observations in each flat cluster have no greater a
cophenetic distance than `t`.
``maxclust`` :
Finds a minimum threshold ``r`` so that
the cophenetic distance between any two original
observations in the same flat cluster is no more than
``r`` and no more than `t` flat clusters are formed.
``monocrit`` :
Forms a flat cluster from a cluster node c
with index i when ``monocrit[j] <= t``.
For example, to threshold on the maximum mean distance
as computed in the inconsistency matrix R with a
threshold of 0.8 do::
MR = maxRstat(Z, R, 3)
cluster(Z, t=0.8, criterion='monocrit', monocrit=MR)
``maxclust_monocrit`` :
Forms a flat cluster from a
non-singleton cluster node ``c`` when ``monocrit[i] <=
r`` for all cluster indices ``i`` below and including
``c``. ``r`` is minimized such that no more than ``t``
flat clusters are formed. monocrit must be
monotonic. For example, to minimize the threshold t on
maximum inconsistency values so that no more than 3 flat
clusters are formed, do::
MI = maxinconsts(Z, R)
cluster(Z, t=3, criterion='maxclust_monocrit', monocrit=MI)
Returns
----------
groups : list of shape (n_clusters, n_members)
Examples
----------
>>> candidate_list = np.array([,...,], shape=[5,3,2]) -> labels of candidate_list = [0,1,0,1,0]
>>> groups = [[0,2,4],[1,3]]
"""
assert isinstance(candidate_list, (np.ndarray, np.generic))
num_of_data, num_time_steps, _ = candidate_list.shape
X = np.array([
candidate_list[i, num_time_steps - 1, :]
for i in range(num_of_data)
])
rads = np.radians(X) # [N,2]
# Clustering with gps-data of 1-time step.
# 'haversine' do clustering using distance transformed from (lat, long)
clusterer = hdbscan.HDBSCAN(min_cluster_size=min_pts,
min_samples=2,
metric='haversine')
labels = clusterer.fit_predict(rads)
print('Before trajectory clustering, labels are ', labels)
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
groups = []
for ulb in range(n_clusters_):
groups.append([])
for i, lb in enumerate(labels):
if lb == -1:
continue
groups[lb].append(i)
total_n_clusters = n_clusters_
# Group refinement considering user's trajectory
for nc in range(n_clusters_):
group_member_mask = (labels == nc)
group_members = candidate_list[group_member_mask]
pdist = tdist.pdist(group_members.transpose([0, 2, 1]),
metric="sspd",
type_d="spherical")
Z = fc.linkage(pdist, method="ward")
sub_labels = sch.fcluster(Z, t, criterion=criterion) - 1
unique_sub_labels = len(set(sub_labels))
if unique_sub_labels == 1:
continue
for ad in range(unique_sub_labels - 1):
groups.append([])
member_indices = list(
compress(range(len(group_member_mask)), group_member_mask))
for sb in range(unique_sub_labels):
sub_group_mask = (sub_labels == sb)
sub_member_indices = list(
compress(range(len(sub_group_mask)), sub_group_mask))
# Noise case
if len(sub_member_indices) == 1:
groups[nc].remove(member_indices[sub_member_indices[0]])
labels[member_indices[sub_member_indices[0]]] = -1
continue
for m in range(len(sub_member_indices)):
# remove from wrong group
groups[nc].remove(member_indices[sub_member_indices[m]])
# add to refined group
groups[total_n_clusters].append(
member_indices[sub_member_indices[m]])
labels[member_indices[
sub_member_indices[m]]] = total_n_clusters
total_n_clusters += 1
print('After trajectory clustering, labels are ', labels)
return groups.copy()
import numpy as np
# Three 2-D Trajectory
traj_A = np.array([[-122.39534, 37.77678],[-122.3992 , 37.77631],[-122.40235, 37.77594],[-122.40553, 37.77848],
[-122.40801, 37.78043],[-122.40837, 37.78066],[-122.41103, 37.78463],[-122.41207, 37.78954],
[-122.41252, 37.79232],[-122.41316, 37.7951 ],[-122.41392, 37.7989 ],[-122.41435, 37.80129],
[-122.41434, 37.80129]])
traj_B = np.array([[-122.39472, 37.77672],[-122.3946 , 37.77679],[-122.39314, 37.77846],[-122.39566, 37.78113],
[-122.39978, 37.78438],[-122.40301, 37.78708],[-122.4048 , 37.78666],[-122.40584, 37.78564],
[-122.40826, 37.78385],[-122.41061, 37.78321],[-122.41252, 37.78299]])
traj_C = np.array([[-122.39542, 37.77665],[-122.3988 , 37.77417],[-122.41042, 37.76944],[-122.41459, 37.77016],
[-122.41462, 37.77013]])
traj_list = [traj_A, traj_B, traj_C]
import traj_dist.distance as tdist
# Simple distance
dist = tdist.sspd(traj_A,traj_B)
print(dist)
# Pairwise distance
pdist = tdist.pdist(traj_list,metric="sspd")
print(pdist)
# Distance between two list of trajectories
cdist = tdist.cdist(traj_list, traj_list,metric="sspd")
print(cdist)