An accurate cardinality estimation using Recursive Neural Networks
Cardinality estimation is a key component in query optimization. To choose the best executing plan, the query optimizer should precisely estimate the selectivity of a SQL predicate -- the number of rows in relation selected by the predicate without actual execution.
python3 -m benchmark.models.treernn.train
python3 -m benchmark.evaluation.eval
If you found this code useful, please cite the following paper: