Simulation in "Trajectory clustering via deep representation learning"
Tested for following versions (Should be compatible for many lower versions):
Tensorflow = 2.2.0
pandas = 1.0.3,
sklearn = 0.22.1
traj_dist = https://github.com/maikol-solis/trajectory_distance (for python2)
traj_dist = https://github.com/THinnerichs/trajectory_distance (for python3)
Generating the synthetic trajectories 'sim_trajectories'
in /simulated_data/
.
Here, we only generate 30 trajectories as the sample.
Embed each trajectory to a fixed-length vector utilizing our framework. Five files is generated in /simulated_data/
sim_trajectories_complete
: Trajectories after attributes completion.sim_trajectories_feas
: Elementary features computed by each pair of continuous recordssim_behavior_sequences
: Behavior sequence generated sliding windows and Feature Extraction Layersim_normal_behavior_sequences
: Moving behavior sequence after normalizaionsim_traj_vec_normal_reverse
: Vector of trajectories generated by Seq2Seq Auto-Encoder Layer
Four distance based trajectory clustering methods(DTW, EDR, LCSS, Hausdorff) were compared with our framwork.
After compution, distance matrices are generated in /distances/
.
Besse P, Guillouet B, Loubes J M, et al. Review and perspective for distance based trajectory clustering[J]. arXiv preprint arXiv:1508.04904, 2015.
"Yao, D., Zhang, C., Zhu, Z., Huang, J., & Bi, J. (2017, May). Trajectory clustering via deep representation learning. In Neural Networks (IJCNN), 2017 International Joint Conference on (pp. 3880-3887). IEEE."
Updated this repo to fit python3 and some other SOTA modules.