time-series-split

Temporal partitioning and backtesting for time-correlated datasets

Cost-aware credit card fraud detection pipeline: time-based split, probability calibration, and business-aligned threshold tuning (AUPRC-first).

Practical guide to validating time-series models: stationary vs non-stationary setups, leakage pitfalls, expanding/rolling TimeSeriesSplit, Ridge/LinearRegression baselines, and bootstrap CIs for metrics. Includes utilities for feature engineering (percent_change) and visualization.

Predicting NBA MVP vote totals across 42 seasons (17,976 player-seasons) with stacked and from-scratch adaptive ensembles — including why the hand-built weighting scheme lost to a plain average

Predicting House Prices — End-to-End Regression Workflow

Predict NYC restaurant inspection grades (A vs B/C+) with leak-safe, time-split ML baselines.

Modelo que predice si un contrato publico peruano terminara siendo modificado, con features construidas sin fuga temporal y particion temporal estricta