Python code for "A unified incremental updating framework of attribute reduction for two-dimensionally time-evolving data". We provide algorithms AAO, DAO, AADO and DAAO for incrementally calculating approximation quality under 4 types of data variation (addition/deletion of object/attribute), and algorithms IARAAO, IARDAO, IARAADO, IARDAAO for incremental feature selection under 4 types of data variation.

Clone this repo.

git clone https://github.com/JsSparkyyx/IAR2DV
cd IAR2DV

Install dependencies by

pip install -r requirements.txt

You can use python src/main.py --exp <exp_name> to run experiment AAO, DAO, IARAAO, IARDAO, IARAADO or IARDAAO on example dataset.

If you want to run experiment on your own dataset, you should first prepare the following file:

• data_name.csv: Each line represents an object, which contains several tokens <attribute 1>,...,<attribute n>,<label> where attribute and label should be a number. You should use \t as the delimiter.

Then you can run experiment on your own dataset by python src/main.py --exp <exp_name> --data <data_path>.