Time-series data is one of the most common forms of data encountered in a wide variety of scenarios such as the stock markets, sensor data, fault monitoring, machine state monitoring, environmental applications, or medical data. The problem of classification finds numerous applications in the time series domain, such as the determination of predefined groups of entities that are most similar to a time series entity whose group is still unknown. Timeseries classification has numerous applications in diverse problem domains:
Financial Markets: In financial markets, the values of the stocks represent time-series which continually vary with time. The classification of a new time series of unknown group can provide numerous insights into descisions about this specific time series.
Medical Data: Different kinds of medical data such as EEG readings are in the form of time series. The classification of such time series, e.g., for a new patient, can provide insights into similar treatment or aid the domain experts in the decisions that have to be made, as similar behavior may indicate similar diseases.
Machine State Monitoring: Numerous forms of machines create sensor data, which provides a continuous idea of the states of these objects. These can be used in order to provide an idea of the underlying behaviors and the groups a time series belongs to.
Spatio-temporal Data: Trajectory data can be considered a form of multi-variate time series data, in which the X- and Y-coordinates of objects correspond to continuously varying series. The behavior in these series can be used in order to determine the class a trajectory belongs to, and then decide, e.g., if a specific trajectory belongs to a pedestrian or to a vehicle.