# AnoFox | Enterprise Forecasting AnoFox is an enterprise forecasting platform for Sales & Operations Planning, delivered as three specialized DuckDB extensions. ## Forecast - [Conformal Prediction](https://anofox.com/docs/forecast/conformal-prediction.md): Generate distribution-free prediction intervals with guaranteed coverage using AnoFox conformal prediction functions. No parametric assumptions required. - [Cross-Validation](https://anofox.com/docs/forecast/cross-validation.md): Backtest forecasts with time-series cross-validation. Create proper train/test splits, forecast on folds, and prevent data leakage in DuckDB. - [Data Quality](https://anofox.com/docs/forecast/data-quality.md): Assess time series data quality across multiple dimensions before forecasting. Score completeness, regularity, stationarity, and length per series. - [Exploratory Data Analysis](https://anofox.com/docs/forecast/eda.md): Profile time series data before forecasting with comprehensive statistical metrics. Compute per-series stats, quality reports, and summary statistics. - [Exogenous Variables](https://anofox.com/docs/forecast/exogenous.md): Incorporate external predictors like temperature, promotions, or pricing into your DuckDB forecasts using ARIMAX, ThetaX, and MFLESX models. - [Features](https://anofox.com/docs/forecast/features.md): Engineer time-series features for analysis and machine learning. Compute 117 tsfresh-compatible statistics in whole-series, rolling, or lagged modes. - [Hierarchy Management](https://anofox.com/docs/forecast/hierarchy.md): Manage hierarchical time series in DuckDB. Combine ID columns, aggregate at hierarchy levels, and reconcile forecasts across regions, stores, and products. - [Installation & Setup](https://anofox.com/docs/forecast/installation.md): Install the AnoFox Forecast DuckDB extension in minutes. Supports Linux, macOS, and Windows on x86_64, ARM64, and WASM architectures. - [Metrics](https://anofox.com/docs/forecast/metrics.md): Evaluate forecast accuracy with 11 built-in metrics including MAE, RMSE, MAPE, MASE, and SMAPE. Compare models and validate predictions in DuckDB. - [Function Finder](https://anofox.com/docs/forecast/overview.md): Browse all AnoFox Forecast functions: EDA, data quality, models, diagnostics, metrics, cross-validation, and conformal prediction. Find the right function fast. ## Forecast - Data-preparation - [Edge Cleaning](https://anofox.com/docs/forecast/data-preparation/edge-cleaning.md): Remove leading and trailing zeros from time series edges to eliminate startup and shutdown artifacts before forecasting in DuckDB. - [Series Filtering](https://anofox.com/docs/forecast/data-preparation/filtering.md): Filter out time series that fail quality criteria. Remove constant, too-short, or excessively gappy series before running forecasts in DuckDB. - [Gap Filling](https://anofox.com/docs/forecast/data-preparation/gap-filling.md): Fill missing timestamps in time series data with NULL values or extend series to a target date. Ensure continuous date ranges before forecasting. - [Imputation](https://anofox.com/docs/forecast/data-preparation/imputation.md): Fill missing values in time series with forward fill, backward fill, linear interpolation, or constant replacement. Prepare clean data for forecasting. ## Forecast - Diagnostics - [Changepoint Detection](https://anofox.com/docs/forecast/diagnostics/changepoints.md): Find structural breaks and regime shifts in time series using Bayesian Online Changepoint Detection. Detect level shifts and trend changes in DuckDB. - [Time Series Decomposition](https://anofox.com/docs/forecast/diagnostics/decomposition.md): Separate time series into trend, seasonal, and residual components using MSTL decomposition. Detrend data and detect peaks in DuckDB. - [Seasonality Detection](https://anofox.com/docs/forecast/diagnostics/seasonality.md): Detect and classify seasonal patterns with 12 detection methods. Identify period lengths and seasonality types for accurate forecasting in DuckDB. ## Forecast - Models - [ARIMA](https://anofox.com/docs/forecast/models/arima.md): Forecast with ARIMA, AutoARIMA, and ARIMAX models in DuckDB. Automatic parameter selection, seasonal differencing, and exogenous variable support. - [Baseline Models](https://anofox.com/docs/forecast/models/baseline.md): Generate benchmark forecasts with Naive, Seasonal Naive, Simple Moving Average, and Random Walk with Drift models. Establish baselines for comparison. - [Exponential Smoothing](https://anofox.com/docs/forecast/models/exponential.md): Forecast with ETS, AutoETS, SES, Holt, and Damped Trend models in DuckDB. Weighted average methods with automatic component selection and optimization. - [Intermittent Demand](https://anofox.com/docs/forecast/models/intermittent.md): Forecast sparse, lumpy demand with Croston, SBA, TSB, and ADIDA models in DuckDB. Specialized methods for inventory planning with many zero observations. - [Multi-Seasonal Models](https://anofox.com/docs/forecast/models/multi-seasonal.md): Handle complex seasonality with TBATS, MSTL, and MFLES models in DuckDB. Capture daily, weekly, and yearly cycles simultaneously in a single forecast. - [Theta Methods](https://anofox.com/docs/forecast/models/theta.md): Forecast with Theta, AutoTheta, OptimizedTheta, and DynamicTheta models in DuckDB. M3 competition-winning decomposition-based forecasting methods. ## Statistics - [Diagnostics](https://anofox.com/docs/statistics/diagnostics.md): Validate regression models with VIF, residual diagnostics, AIC, and BIC. Check assumptions, detect multicollinearity, and compare models in DuckDB. - [Installation & Setup](https://anofox.com/docs/statistics/installation.md): Install the AnoFox Statistics DuckDB extension in minutes. Supports Linux, macOS, and Windows on x86_64, ARM64, and WASM architectures. - [Function Finder](https://anofox.com/docs/statistics/overview.md): Browse all AnoFox Statistics functions: regression models, hypothesis tests, diagnostics, and demand analysis. Find the right SQL function for your task. - [DuckDB Patterns](https://anofox.com/docs/statistics/utilities.md): Learn how AnoFox Statistics integrates with DuckDB SQL patterns. Use aggregate functions with GROUP BY, window functions with OVER, and batch predictions. ## Statistics - Demand - [Demand Classification](https://anofox.com/docs/statistics/demand/classification.md): Classify demand as regular or intermittent, identify best-fit distributions, and detect anomaly patterns with AID demand classification in DuckDB. ## Statistics - Hypothesis-tests - [Categorical Tests](https://anofox.com/docs/statistics/hypothesis-tests/categorical.md): Run chi-square, Fisher's exact, and goodness-of-fit tests for categorical data directly in DuckDB. Analyze independence and distributions with pure SQL. - [Correlation Tests](https://anofox.com/docs/statistics/hypothesis-tests/correlation.md): Measure and test correlations with Pearson, Spearman, and Kendall methods in DuckDB. Compute significance tests and confidence intervals for relationships. - [Forecast Model Selection](https://anofox.com/docs/statistics/hypothesis-tests/forecast-evaluation.md): Compare forecast accuracy between competing models using the Diebold-Mariano test. Determine statistically significant differences in predictive performance. - [Nonparametric Tests](https://anofox.com/docs/statistics/hypothesis-tests/nonparametric.md): Run distribution-free statistical tests in DuckDB: Mann-Whitney U, Kruskal-Wallis, Wilcoxon signed-rank, and Friedman tests. No normality assumptions required. - [Parametric Tests](https://anofox.com/docs/statistics/hypothesis-tests/parametric.md): Run parametric hypothesis tests in DuckDB: Shapiro-Wilk normality, one-sample and two-sample t-tests, paired t-test, one-way ANOVA, and Levene's test. ## Statistics - Regression - [Constrained Models](https://anofox.com/docs/statistics/regression/constrained.md): Fit constrained regression models in DuckDB with Bounded Least Squares and Non-Negative Least Squares. Enforce coefficient bounds and positivity constraints. - [Generalized Linear Models](https://anofox.com/docs/statistics/regression/generalized.md): Fit Poisson GLM for count data and ALM with 24 error distributions in DuckDB. Model non-Gaussian responses for defects, arrivals, and event counts. - [Linear Models](https://anofox.com/docs/statistics/regression/linear.md): Fit OLS, WLS, and RLS regression models in DuckDB with full inference: coefficients, p-values, R-squared, confidence intervals, and residual diagnostics. - [Regularized Models](https://anofox.com/docs/statistics/regression/regularized.md): Handle multicollinearity and high-dimensional data with Ridge and Elastic Net regression in DuckDB. L2 regularization with automatic penalty selection. ## Tabular - [Anomaly](https://anofox.com/docs/tabular/anomaly.md): Detect outliers with Z-Score, IQR, Isolation Forest, DBSCAN, and explainable OutlierTree methods in DuckDB. Statistical and ML-based anomaly detection. - [Financial](https://anofox.com/docs/tabular/financial.md): Validate monetary amounts, format currencies, convert between ISO currency codes, and calculate percentages with AnoFox Tabular financial functions in DuckDB. - [Installation](https://anofox.com/docs/tabular/installation.md): Install the AnoFox Tabular DuckDB extension from the community registry or build from source. Supports Linux, macOS, and Windows. - [Operations](https://anofox.com/docs/tabular/operations.md): Compare and diff datasets in DuckDB with hash-based and join-based table comparison functions. Detect row-level changes between data snapshots. - [Function Finder](https://anofox.com/docs/tabular/overview.md): Browse all 78 AnoFox Tabular functions: validation, financial, quality metrics, anomaly detection, PII detection, and dataset comparison operations. - [PII Detection](https://anofox.com/docs/tabular/pii.md): Detect and mask 17 types of personally identifiable information in DuckDB: SSN, credit cards, IBAN, names, emails, and more. Row-level and table-level scanning. - [Quality](https://anofox.com/docs/tabular/quality.md): Measure data quality with nullness, distinctness, freshness, volume, consistency, and schema match functions. Profile your data directly in DuckDB. - [Validation](https://anofox.com/docs/tabular/validation.md): Validate emails, phone numbers, VAT IDs, addresses, and postal codes directly in DuckDB. RFC 5322, DNS, and SMTP checks with country-specific rules. ## Workflow ## Additional Resources