rel-avito: user visits (Classification)

This example now materializes the official rel-avito tables through the relbench Python library and then runs the user-visits GraphReduce task against those local files.

from relbench_avito_user_visits import run_rel_avito_user_visits

df, auc, n_features, materialized, target = run_rel_avito_user_visits()
print(materialized)
print(target, auc, n_features, len(df))

Current implementation:

• examples/relbench_avito_user_visits.py
• examples/relbench_avito_common.py

Latest Tested Result

Latest local run from tests/data/relbench/run_reports/relbench_results.json:

• test ROC AUC: 0.8299
• test average precision: 0.9655
• feature count: 281
• rows: train 86619, validation 29979, test 36129
• runtime: 25.3s (2026-04-09T22:25:04Z)

Interactive Runner

Save API URL Run rel-avito user visits

Idle