← Back
Home › Systems › AI applications

demand-forecasting

demand forecasting

Amazon forecasting

SystemTechniques · 7Primitives · 29 demand-forecasting Gradient-boosted decision trees (GB…machine-learning · medium: Add small regression trees one at a time, each fitted to the loss gradient; histogram splits over 256 bins make it the strongest tabular default.›Arimastatistics · medium: ARIMA/SARIMA forecasting: differencing, AR and MA terms, order selection, maximum-likelihood fitDeep forecasting modelsstatistics · medium: Neural forecasters (DeepAR, N-BEATS, TFT, patch transformers) with probabilistic outputs across many seriesHierarchical reconciliationstatistics · medium: Making forecasts coherent across hierarchy levels (bottom-up, top-down, MinT)Columnar storagestorage · medium: Store each column contiguously so analytic scans read only the columns they need, compress well and run as tight SIMD loops.›Vectorized executionstorage · medium: Query operators pass batches of about 1-2K values per column instead of single tuples, amortizing interpretation overhead and enabling tight SIMD loops.›Rollup aggregationstorage · medium: Fold raw events into one mergeable partial aggregate per time bucket and dimension key, then merge fine buckets into coarser ones, so long ranges read few rows and stay exact.›Quantile regressionmachine-learning · low: Predicting quantiles with the pinball loss for probabilistic forecasts and prediction intervalsTime-series cross-validation (rolli…statistics · low: Estimate forecast error by repeatedly training on the past and testing on the next h steps, moving the origin forward; never trains on the future.›Evaluation metricsmachine-learning · low: Precision, recall, ROC AUC, average precision, log loss, NDCG and pass@k computed correctly, plus paired confidence intervals for comparing models.›Linear and logistic regressionmachine-learning · low: Fit a weighted sum of features to a number (least squares) or a probability (logistic); the baseline every model must beat.›Croston methodstatistics · low: Forecasting intermittent demand by separately smoothing demand sizes and intervals (Croston, SBA, TSB)Exponential smoothing (ETS, Holt-Wi…statistics · low: Forecast by keeping a few running states (level, trend, seasonal) updated by a fraction of each forecast error; O(1) memory per series and a strong baseline.›STL decomposition (seasonal-trend w…statistics · low: Split a series into trend, seasonal and remainder with iterated loess smoothing; the seasonal shape may drift and robust weights keep outliers out of it.›Quantile sketchesdata-structures · low: Mergeable quantile sketches: t-digest, DDSketch, HDR histogramsLoss functionsmachine-learning · low: The scalar a model minimizes: cross-entropy for classes and tokens, MSE or Huber for values, each with a stable form and simple gradient.›Regularizationmachine-learning · low: Techniques that trade training fit for generalization: weight decay, dropout, early stopping, label smoothing and data mixing, with the exact update rules and defaults.›Finite differencesnumerics · low: Replace derivatives with weighted differences of grid values; simple stencils plus a stability condition (CFL) turn PDEs like heat, advection and waves into array loops.›L-BFGSnumerics · low: Quasi-Newton minimizer that approximates the inverse Hessian from the last m gradient differences via the two-loop recursion; O(m·n) memory and work per step.›Kalman filterrobotics · low: Optimal recursive estimator for a linear system with Gaussian noise: fuses a motion model with noisy measurements into a state estimate plus its covariance.›Recurrent networksmachine-learning · low: GRU and LSTM cells for streaming sequence modelsTransformer blockml-training · low: The decoder-only pre-norm block (attention plus MLP on a residual stream) behind every modern LLM, with exact shapes, parameter, FLOP and KV-cache counts.›LU and Cholesky factorizationnumerics · low: Factor A = PLU (any square matrix, partial pivoting) or A = LLᵀ (SPD) once in O(n³), then solve each right-hand side in O(n²).›+ 13 more primitives

Similar systems

Share the most building blocks with this one.

retail-merchandising 21 %metric-anomaly-detection 19 %credit-scoring 18 %test-selection 16 %

In-house systems

Named systems of this kind at large companies, with their paper or write-up.

SCOT forecastingAmazon