Introduction
A time series is a sequence of values measured over time: daily website visitors, hourly electricity demand, a patient's heart rate, the number of passengers on a bus line each day. Forecasting means predicting future values from past ones. It's one of the most common business uses of machine learning: planning staff, ordering stock, setting prices.
Think of a weather forecaster. Before using a fancy model, they know the simplest predictions are surprisingly good: "tomorrow will be like today", or "this July will be like last July". Any clever model must beat those simple guesses to be worth using.