trajectory2vec

Simulation in "Trajectory clustering via deep representation learning"

Required Packages:

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)

Useage:

simulate_data.py:

Generating the synthetic trajectories 'sim_trajectories' in /simulated_data/. Here, we only generate 30 trajectories as the sample.

tmb2vec.py:

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 records
sim_behavior_sequences : Behavior sequence generated sliding windows and Feature Extraction Layer
sim_normal_behavior_sequences : Moving behavior sequence after normalizaion
sim_traj_vec_normal_reverse : Vector of trajectories generated by Seq2Seq Auto-Encoder Layer

compared_methods.py:

Four distance based trajectory clustering methods(DTW, EDR, LCSS, Hausdorff) were compared with our framwork. After compution, distance matrices are generated in /distances/.

Reference:

Besse P, Guillouet B, Loubes J M, et al. Review and perspective for distance based trajectory clustering[J]. arXiv preprint arXiv:1508.04904, 2015.

Citing trajectory2vec

"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."

Update

Updated this repo to fit python3 and some other SOTA modules.

THinnerichs / trajectory2vec