erssss / TDFMLPS

• This repository is for paper High Precision $\neq$ High Cost: Temporal Data Fusion for Multiple Low-Precision Sensors.

File Structure

• code: source code of algorithms.
  • The functions have the same names as the algorithms in the paper.
  • As for TruthFinder, Sums, Investment, CRH, GATD, we use the open source implemenations for them, i.e.,
    • TruthFinder, Sums, Investment: https://github.com/joesingo/truthdiscovery
    • CRH, GATD: https://sites.google.com/iastate.edu/qili
• data: dataset source files used in experiments.
• full.pdf: the full version of our manuscript, including all the complete proofs in appendix.

Dataset

• GPS
  • Manual collection
  • Format: timestamp(1),observations from different sensors(2-5), true value(6)
• WEATHER
  • https://www.aerisweather.com
  • https://www.worldweatheronline.com/
  • https://www.wunderground.com/
  • run gen_weather.py with the ID get from the official website and parameters specified in the code.
• IMU: https://github.com/dusan-nemec/mems-calib
• GINS: https://github.com/i2Nav-WHU/awesome-gins-datasets

Dependencies

numpy==1.24.2
pandas==1.4.3
pyclustering==0.10.1.2
scikit_learn==1.1.1

Instruction

To run the program, use the following command-line arguments:

• --dataname: Input file name (default is "GPS").
• --alg: Method name (default is "DFDP").
• --k: The parameter kappa (default is 3).

eg.

main.py --data GPS --alg DFRC --k 3

or you can run directly:

cd code
python main.py