This is a seed guided neural metric learning approach for calculating trajectory similarities.

Pytorch, Numpy, trajectory_distance

Please create 3 empty folders:

*data: Path of the original data which is organized to a trajectory list. Each trajectory in it is a list of coordinate tuples (lon, lat).

*features: This folder contains the features that generated after the preprocessing.py. It contains four files: coor_seq, grid_seq, index_seq and seed_distance.

*model: It is used for placing the NeuTraj model of each training epoch.

Due to the file limit of Github, we put the dataset on other sites. Please first download the data and put it in data folder. The toy dataset can be download at: https://www.dropbox.com/s/ejoo1j21vjq7t7a/toy_trajs?dl=0

Run preprocessing.py. It filters the original data and maps the coordinates to grids. After such process, intermediate files which contain coor_seq, grid_seq, and index_seq are generated. Then, we calculate the pair-wise distance under the distance measure and get the seed_distance.

Run train.py. It trains NeuTraj under the supervision of seed distance. The parameters of NeuTraj can be modified in /tools/config.py