Code for DASFAA 2024 "Flexible Contact Correlation Learning on Spatio-Temporal Trajectories"
- torch
- torch_geometric
- networkx
- geopy
- numba
- transbigdata
- Step1: Download data from https://www.cs.rutgers.edu/~dz220/data.html.
- Step2: Put the data file in ../datasets/sz/, and unzip it as TaxiData.txt.
- Step3: Run
mkdir -p data/sz
python preprocess_sz.py
mkdir data/sz/s_t temp_data
python ground_truth.py --dataset sz --contact_factor 3 --score_type s_t
Here, contact_factor and score_type are introduced in the Preliminaries section.
mkdir logs models_pth
python pretrain.py --dataset sz
python main.py --dataset sz --contact_factor 3 --score_type s_t --lr 3e-4 --mode train
python main.py --dataset sz --contact_factor 3 --score_type s_t --mode test
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent