A time series is a set of observations ordered in time usually collected at regular intervals.
Proposed by @Al_Grigor
Card 1 of 259. Answer: A time series is a set of observations ordered in time usually collected at regular intervals.
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
A time series is a sequence of observations recorded in time order, usually at equal intervals such as hourly, daily or monthly. Order carries information: each value tends to depend on earlier ones, so unlike rows in most machine learning datasets, observations are not independent and cannot be shuffled.
Analysts split a series into components. Monthly electricity demand might show a trend (rising with population), seasonality (summer and winter peaks repeating every 12 months), economic cycles and random noise. A standard baseline is the seasonal naive forecast: each month's value from a year earlier. To use an ordinary regressor, build lag features, such as the values 1 and 12 months back.
Follow-ups concern validation and stationarity. Avoid shuffled k-fold cross-validation; train on the past and test on the future. ARMA models assume a stationary series, whose mean and variance do not change over time; ARIMA differences it first to remove trend. And 'usually' matters: trades and sensor events arrive at irregular times and are often resampled to a regular grid.
