AI Ml Timeseries

Time Series Forecasting - Production Patterns

Scope note: This skill covers forecasting system construction and evaluation. It is not part of the LLM-build or LLM-training stack — route LLM lifecycle, prompting, or provider questions to ai-llm.

July 2026 posture: define a cutoff timestamp before modelling, start with strong baselines, prefer horizon-aware validation over IID thinking, treat known-future covariates explicitly, and verify fast-moving tooling against current official docs before recommending it.

This skill is the implementation guide for forecasting systems:

• timestamp integrity, frequency checks, and point-in-time feature design
• local, global/panel, and hierarchical forecasting workflows
• leakage-safe backtesting, horizon-wise evaluation, and business-loss alignment
• probabilistic forecasting, calibration, and interval quality
• time-series foundation models (TSFMs) and zero-shot benchmark patterns
• forecasting-specific handoff, fallback, and lineage requirements

Use this skill for forecasting depth. Use sibling skills for general data science, generic LLM strategy, or full production operations.

When To Use This Skill

Activate this skill when the user asks for:

• building or reviewing a forecast model
• choosing between local, global/panel, hierarchical, or foundation-model approaches
• creating lag, rolling, calendar, or known-future covariate features
• designing a rolling-origin backtest or fixing temporal leakage
• selecting forecasting metrics by horizon, segment, or business cost
• adding prediction intervals, quantiles, or conformal calibration
• comparing Chronos-2, Chronos-Bolt, Toto, TimesFM, AutoGluon TimeSeries, MLForecast, skforecast, or classical baselines
• defining forecast-specific fallback, lineage, and handoff requirements

Scope Boundaries

• General EDA, tabular modelling, experiment design, or reusable DS workflow -> ai-ml-data-science
• Deployment architecture, monitoring stack, release gates, incident playbooks -> ai-mlops
• Generic LLM lifecycle, prompting, or provider selection -> ai-llm
• RAG and search systems -> ai-rag

Keep this skill focused on forecast creation, validation, and forecasting-specific operational patterns.

ASCII Flow

forecast need
  |
  v
forecast contract
  target + horizon + cutoff + granularity + business loss + update cadence
  |
  v
time-series validation
  timestamp integrity + leakage checks + rolling-origin backtest
  |
  v
model pattern
  naive baseline -> local/classical -> feature ML -> global/panel -> TS foundation
  |
  v
forecast handoff
  point/interval forecasts + lineage + fallback + monitoring trigger

Quick Reference

Workflow

1. Define the forecast target, horizon, cutoff timestamp, granularity, and business-loss shape first.
2. Route generic DS workflow or full production-ops work to the adjacent skill when forecasting depth is not the main need.
3. Choose the baseline, feature pattern, and model family from the decision tree.
4. Run leakage-safe backtests, compare horizon-aware metrics, and add interval or calibration checks when decisions need uncertainty.
5. Verify fast-moving tooling and TSFM claims with the navigation and fact-checking sources before final recommendations.

Current-Facts Protocol For Current-Best-Tooling Recommendations

Use this protocol whenever the user asks for the best, latest, current, or still relevant forecasting tooling — TSFM releases, benchmark leaderboards, and package APIs move on a roughly quarterly cadence, so treat any specific version number or benchmark figure below as provisional pending a live check:

1. Start from data/sources.json.
2. Verify fast-moving facts with official docs or official repositories.
3. Separate stable guidance from volatile facts such as current model capabilities, package APIs, or hosted deployment options.
4. Recommend a default plus 1-2 alternatives, with tradeoffs by horizon, covariates, scale, and operational cost.

Use this protocol for questions such as:

• "Should I use MLForecast, skforecast, or AutoGluon TimeSeries?"
• "Should I use Chronos-2, Chronos-Bolt, or Toto as my zero-shot baseline?"
• "What is the current MAPIE API for time-series conformal intervals?"
• "Should I still be using Prophet or TBATS here?"

Decision Tree: Choose The Forecasting Pattern

Need to build or review a forecast:
    ├─ One series or a small local portfolio?
    │   ├─ Few covariates, interpretability first -> naive/seasonal naive + ETS/SARIMAX baseline
    │   └─ Nonlinear effects or richer covariates -> feature-based boosting
    │
    ├─ Many related series?
    │   ├─ Need one shared model with covariates -> global/panel forecasting
    │   └─ Need rollups to add up across levels -> hierarchical forecasting + reconciliation
    │
    ├─ Intermittent or zero-heavy demand?
    │   ├─ Very sparse and operationally simple -> Croston/SBA/ADIDA baseline
    │   └─ Need richer covariates or scale -> global boosting with zero-aware features
    │
    ├─ Need uncertainty, service levels, or inventory decisions?
    │   └─ Add quantiles, conformal intervals, or distributional forecasts
    │
    ├─ Long horizon or low-feature setup?
    │   ├─ Need strong zero-shot baseline -> TS foundation models
    │   └─ Need supervised optimization on many series -> global/deep forecasting
    │
    └─ Need production handoff?
        └─ Record cutoff timestamp, feature contract, fallback, lineage, and retraining trigger

Core Principles

1. Cutoff Timestamp Before Features

• Define the exact prediction timestamp before building labels or features.
• Every lag, rolling aggregate, calendar flag, and external signal must be justified relative to what is known at that cutoff.
• Known-future covariates and future-unknown covariates must be treated differently.

2. Baselines First

• Always compare against naive and seasonal-naive baselines.
• Add ETS/SARIMAX or other local classical baselines when interpretability matters.
• Candidate models do not earn deployment if they barely beat a simple baseline.

3. Horizon And Slice Evaluation

• Report accuracy by horizon, not only a single global score.
• Slice by segment, geography, SKU family, volume band, or any business-relevant cohort.
• Prefer MASE/WAPE/MAE over MAPE when zeros or near-zeros exist.

4. Global/Panel Before Per-Series Complexity

• When many related series exist, default to a shared global/panel approach before building separate bespoke models.
• Use hierarchical reconciliation when forecasts must remain coherent across levels.
• Promote complexity only when it earns accuracy, calibration, or operational simplicity.

5. Probabilistic When The Decision Is Risk-Sensitive

• Use quantiles, conformal intervals, or full predictive distributions when downstream actions depend on uncertainty.
• Evaluate both coverage and sharpness; wide intervals with nominal coverage are not automatically useful.
• Reassess calibration after every retrain or major model change.

6. Forecasting-Specific Handoff

• A forecast package is incomplete without cutoff timestamps, horizon definition, feature contract, metric definitions, fallback rules, and lineage metadata.
• Keep forecasting-specific handoff guidance here; route full deployment architecture to ai-mlops.

Known Traps

• Mixing future-known covariates and future-unknown covariates in the same feature path without documenting which values are actually available at forecast time.
• Using one global backtest score to justify deployment when horizon-specific error behavior differs materially.
• Using MAPE on zero-heavy, intermittent, or near-zero series and then comparing models on unstable percentages.
• Building bespoke per-series models before testing a strong global or panel baseline on related series.
• Ignoring hierarchy and coherence when downstream consumers expect rollups to add up.

Common Anti-Patterns

• Reusing IID validation habits from generic tabular ML instead of rolling-origin or expanding-window evaluation.
• Treating decomposition visuals as evidence of production signal without backtesting the actual decision horizon.
• Using feature-rich models whose future covariates are unavailable or operationally too expensive to maintain.
• Comparing TS foundation models to weak baselines and calling the result strategic proof.

Navigation: Core References

Data Integrity And Features

• TS EDA Best Practices - Timestamp integrity, missingness, decomposition, and stability checks
• Lag & Rolling Patterns - Leakage-safe lags, rolling windows, and calendar patterns
• Global & Panel Forecasting Patterns - Shared models, panel schemas, known-future covariates, grouped evaluation

Model And Strategy Selection

• Model Selection Guide - May 2026 model-family decision matrix
• LightGBM TS Patterns - Feature-based/global boosting patterns, MLForecast/skforecast workflows
• Multi-Step Forecasting Patterns - Direct, recursive, and seq2seq tradeoffs
• Intermittent Demand Patterns - Sparse-demand baselines and zero-aware modelling

Validation, Uncertainty, And Advanced Forecasting

• Backtesting Patterns - Rolling-origin evaluation, panel-aware backtests, and metric design
• Probabilistic Forecasting - Quantiles, conformal methods, calibration, and scoring rules
• Hierarchical Forecasting - Coherent forecasts and reconciliation methods
• Time-Series Foundation Model Patterns - Chronos, TimesFM, zero-shot baselines, and TSFM usage
• Anomaly Detection Patterns - Residual and interval-based anomaly workflows

Handoff And Forecast Operations

• Forecast Governance Patterns - Cutoff timestamps, lineage, fallback rules, and forecast contracts
• Production Deployment Patterns - Broader production controls; use when forecasting-specific guidance must connect to deployment

Templates

Data Preparation

• TS EDA Template - Reproducible structure for timestamp and seasonality review
• Resample & Fill Template - Resampling, gap rules, and fill policies

Feature And Model Design

• Lag & Rolling Features - Leakage-safe feature specification
• Calendar Features - Known-future business calendar and event feature spec
• Forecast Model Template - Forecast package contract for local, global, or hierarchical models
• Multi-Step Strategy - Direct, recursive, and seq2seq strategy contract

Evaluation And Uncertainty

• Backtest Template - Rolling-origin or expanding-window evaluation spec
• TS Metrics Template - Horizon, slice, business-loss, and probabilistic metric contract

Foundation Models

• TS Foundation Model Template - Zero-shot TSFM benchmark and evaluation scaffold

Scripts

# Horizon-wise accuracy: MAE, RMSE, MAPE, skill score vs naive baseline
python scripts/ts_evaluator.py backtest --input data/sample-forecast-results.json

# Probabilistic calibration check for 50%, 80%, 90% prediction intervals
python scripts/ts_evaluator.py calibration --input data/sample-forecast-results.json

# Full Markdown evaluation report written to file
python scripts/ts_evaluator.py report --input data/sample-forecast-results.json --output /tmp/ts-eval-report.md

Data Files

External Sources

See data/sources.json for current primary sources across:

• classical forecasting references
• official docs for MLForecast, HierarchicalForecast, skforecast, AutoGluon TimeSeries, LightGBM, and MAPIE
• official Chronos-2/Chronos-Bolt, TimesFM, and Toto repositories
• fev-bench benchmark and governance references used for high-impact deployments

Related Skills

• ai-ml-data-science - General DS workflows, experiment design, and broader modelling patterns
• ai-mlops - Deployment architecture, monitoring, and release operations
• ai-llm - Provider/model lifecycle questions outside time-series forecasting
• data-sql-optimization - Storage and query design for time-series marts

Fact-Checking

• Known bugs, regressions, framework/compiler/runtime footguns, and version-specific crash or workaround guidance must be verified against current primary web sources before being treated as current fact.
• Verify volatile external facts, package APIs, hosted options, and model capabilities before final answers.
• Prefer official docs, official repositories, standards, or release notes.
• If web access is unavailable, state that clearly and mark time-sensitive guidance as unverified.

Learnings Loop

Before applying this skill on a non-trivial task, read learnings.consolidated.md in this directory (and learnings.md if present).

After applying it, if you encountered a pattern worth remembering, a mistake worth preventing, or a domain fact that surprised you, append one dated bullet to learnings.md via agents-skills-feedback-loop/scripts/append_learning.py. Do not modify SKILL.md itself.