Family
XGBoost + LightGBM + ARIMA ensemble
in-developmentMITtabularretail-facing
Spending Predictor.
Time-series spending forecast (7d, 30d, 90d horizons) — XGBoost + LightGBM + ARIMA ensemble
Foundation · XGBoost + LightGBM + ARIMA ensemble
What it does
The honest description.
Time-series spending forecast (7d, 30d, 90d horizons) — XGBoost + LightGBM + ARIMA ensemble
Architecture
How it works.
User monthly spending history (min 3 months) aggregated by category, temporal features (seasonality, trend), income baseline. 7-day / 30-day / 90-day spending forecasts per category with MAPE error bounds (target < 25%) and ensemble confidence.
Training data & drift
What it learned from, and how we watch it.
Source
Inputs: User monthly spending history (min 3 months) aggregated by category, temporal features (seasonality, trend), income baseline.
Drift baseline · current quarter
Drift baseline not yet documented for this model.
What it's NOT
- XGBoost + LightGBM + ARIMA ensemble with Bayesian weight optimization.
- Walk-forward validation documented but current code uses simple time-split.
- Self-healing via TabularStrategy (D-DRIFT-1).
Who calls this model
Across the house and the API.
Internal routes
No public routes documented yet.
Featured in
/features/goals/features/spending-budgetsVersion 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.