JerryLead / Table-SGD

Table-SGD

Run

python run/sgd.py <path to dataset> --project=<wandb project name> --dataset=<dataset name>

Results

We conducted tests on these three datasets using our Table-SGD and compared the results with baseline GD [1, 2, 3, 4].

1. MIMIC-III
2. Yelp
3. MovieLens-1M

We obtained the original dataset from the official site and performed some preprocessing to convert all data into a numerical format that can be easily processed by the model.

Reference

1. Arun Kumar et al: Learning Generalized Linear Models Over Normalized Data. SIGMOD 2015
2. Maximilian Schleich et al: Learning Linear Regression Models over Factorized Joins. SIGMOD 2016
3. Lingjiao Chen, et al: Towards Linear Algebra over Normalized Data. VLDB 2017
4. Maximilian Schleich et al: A Layered Aggregate Engine for Analytics Workloads. SIGMOD 2019