Family
LightGBM (binary)
productionMITtabularinternal
P2P / P2M Classifier.
Peer-to-peer vs peer-to-merchant classifier. PR #14: retain (consumer exists). Followup: ml-tabular Q3 trainer or deprecate.
Foundation · LightGBM (binary)
What it does
The honest description.
Peer-to-peer vs peer-to-merchant classifier. PR #14: retain (consumer exists). Followup: ml-tabular Q3 trainer or deprecate.
Architecture
How it works.
amount, payment_method_score, counterparty_in_merchant_directory, counterparty_account_age_days, counterparty_prior_transactions, is_recurring_payee. P(is_p2m) ∈ [0, 1]; threshold at 0.5 for binary P2P vs P2M. Used downstream for fee determination + transaction routing + as a cross-model feature into fraud_detection.
Training data & drift
What it learned from, and how we watch it.
Source
Inputs: amount, payment_method_score, counterparty_in_merchant_directory, counterparty_account_age_days, counterparty_prior_transactions, is_recurring_payee.
Drift baseline · current quarter
Drift baseline not yet documented for this model.
What it's NOT
- PR4 (B3): wired into data_sync.MODEL_TO_UPLOAD_TYPE (UploadType.labels, target is_p2m) for label-aware training; the labels-parquet dispatch is DORMANT until a labeled corpus lands and served_mode stays heuristic.
Version history · immutable once published
Every release, traceable.
Data Mukodi holds about you is subject to DPDP §12 erasure (use /consent in the app), and any concern about this model can be raised via [email protected] under DPDP §13 grievance, or via /security.