Forecasting
Forecasting is the process of predicting future values or outcomes from information currently available, often including historical observations.
Outlier Detection in Python
The process of predicting future values or events from historical or current data.
- How it is used
- Uses expected future values and forecast errors to identify point anomalies in time-series data.
- How deep it goes
- Explains forecast errors, robust scoring, confidence ranges, model fitting, evaluation, and horizon limitations without deriving forecasting theory.
- Examples
- Demonstrates synthetic weekly data, ARIMA and Random Forest forecasts, an injected outlier, MAPE evaluation, and UPS Twitter volume with Prophet.
- What you can do after
- Apply forecasting with Darts, scikit-learn, or Prophet to estimate values, measure errors, and flag unusually large deviations.
Demand Forecasting Best Practices
A forecast is a prediction about a future quantity or condition based on available information.
- How it is used
- Supports supply-chain production, inventory, product, and cash-flow decisions at decision-specific horizons and granularities.
- How deep it goes
- Explains objective setting, hierarchy trade-offs, error evaluation, time-series components, and model selection without implementation detail.
- Examples
- Uses role-based planning decisions, product-family versus SKU forecasts, seasonal quarterly demand, and a 36.1% MAE forecast.
- What you can do after
- Evaluate/understand which forecast granularity and method fit available data, decision needs, interpretability, and accuracy trade-offs.
Read how Demand Forecasting Best Practices explains forecasting →
Time Series Forecasting Using Foundation Models
Time-series forecasting is the task of predicting future values of a variable or collection of variables indexed over time.
- How it is used
- Deploys reusable or adapted models to predict future time-indexed values across forecasting, anomaly detection, and classification workflows.
- How deep it goes
- Explains zero-shot, prompt-based, and reprogrammed workflows, but stops short of a universally optimal model or complete benchmark.
- Examples
- Uses weekly store sales, daily blog visits, PromptCast, and models including TimeGPT, Chronos, Lag-Llama, Moirai, TimesFM, and Time-LLM.
- What you can do after
- Apply it via foundation models or LLM prompts, compare speed and performance, and select methods according to data, horizon, hardware, and resources.
Read how Time Series Forecasting Using Foundation Models explains forecasting →
Related concepts
- arimaincludes · ARIMA is a statistical time-series model that combines autoregressive terms, differencing, and moving-average terms to represent temporal dependence and generate forecasts.
- statistical modelingproduced by · Statistical modeling is the use of mathematical or probabilistic models to represent patterns, relationships, or variability in data for description, inference, or prediction.
- Time-LLMused by · Time-LLM is a language-model-based framework that adapts existing large language models for time-series forecasting.
- demand forecastincludes · A demand forecast is an estimate or prediction of the quantity of a product, service, or resource that will be needed during a specified future period.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.