First, Install the anaconda. Then, create an virtual python environment,
conda env create -f environment.yml
Activate the environment, conda activate EBRanker
Please download datasets and you need to configure LTRlocal_dataset_info.txt to specify where you put those datasets.
With the setting.json generated from above cmd, you can run the following to run a batch of scripts.
python main.py --progressbar=false --rankListLength=5 --query_least_size=5 --positionBiasSeverity=1 --n_iteration=57505 --NewItemEnterProb=1.0 --dataset_name=MQ2007 --ExpandFeature=True --Ranker=NNTopK --random_seed=3 --log_dir=localOutput
You can choose different datasets, Ranke, ExpandFeature or not.