AnyGrad

Approximation of large dependency matrices in an 'anytime' fashion using gradient ascent

Execute

The easy way:

sbt assembly
bash run-experiment.sh -t target-dir -f filepath -p parallel [0, 1] -e experiment [t-over-m] -r repetitions [Int]

• t: Where to store the results. Results are exported as json and stored in a .txt file.
• f: Path to the data (csv-file, delimiter=",")
• p: Use parallelization
• e: Select the desired experiment. Possible options are t-over-m
• r: Specify the number of repetitions of the experiment