Time Series
A data storage approach that retains sampling points arranged along a time axis, allowing rate-related metrics such as differences, slopes and moving averages to be computed after the fact, as opposed to the point-wise reduction of aggregate counters.
Timeline (last 90 days)
The essential difference between aggregate counters and time series is that aggregate counters perform an irreversible reduction when an event occurs (summing, taking the maximum, updating the timestamp), storing the result and discarding the raw event, whereas time series retain time-ordered sample points, allowing differences, slopes and moving averages to be computed afterwards.
Time series data exhibit temporal dependencies, trends, and autocorrelation structures, and therefore almost inevitably violate the exchangeability assumption.
The authors instantiate the framework across three data domains: binary images, tabular data, and time series.
All Facts (3)
The essential difference between aggregate counters and time series is that aggregate counters perform an irreversible reduction when an event occurs (summing, taking the maximum, updating the timestamp), storing the result and discarding the raw event, whereas time series retain time-ordered sample points, allowing differences, slopes and moving averages to be computed afterwards.
50%UnverifiedTime series data exhibit temporal dependencies, trends, and autocorrelation structures, and therefore almost inevitably violate the exchangeability assumption.
50%UnverifiedThe authors instantiate the framework across three data domains: binary images, tabular data, and time series.
50%