STAIR and L-STAIR

This is the implementaion for the paper Interpretable Outlier Summarization

Requirements

python 3.7.0
numpy 1.19.4
scikit-learn 1.2.2
tqdm 4.65.0

Datasets

All the datasets are provided here. you can download them and move them into the folder data. The content under the folder data should be:

-- data
    -- SpamBase_withoutdupl_norm_40.arff
    -- Pima_withoutdupl_norm_35.arff
    -- cover.mat
    -- mammography.mat
    -- PageBlocks_norm_10.arff
    -- pendigits.mat
    -- satellite.mat
    -- satimate-2.mat
    -- shuttle.mat
    -- Thursday-01-03-2018_processed.csv
    -- winequality-white.csv

Cammands

Outlier Detection Task

For Outlier Detection datasets, you need to first enter the OutlierDetection folder:

cd OutlierDetection

The following command will run the baselines ID3, CART and our algorithm STAIR successively on the dataset PageBlock.

python main.py PageBlock

If you need to run the algorithm L-STAIR on the dataset PageBlock, you can use the following command:

python lstair_main.py PageBlock

If you need to run the algorithms on other dataset, simply change the dataset name PageBlock into other names such as Pendigits, Pima and so on.

MultiClass Classification Task

For Classification datasets, you need to first enter the MultiClassClassification folder:

cd MultiClassClassification

The following command will run the baselines ID3, CART and our algorithm STAIR successively on the dataset Wine.

python main.py Wine

The command to run L-STAIR on the dataset Wine is:

python lstair_main.py Wine