Glossary¶
Time Series¶
A time series is a sequence of data points collected or recorded over time to track the behavior or measurement of a specific subject or variable across time, allowing analysts to observe trends, patterns and changes.
Signal¶
A Signal is a time varying sequence of values recorded at arbitrary intervals,
where these values can be compared across timestamps. A Signal comprises a
series of <timestamp, value>pairs that represent the same quantity over time.
Numerical Signal¶
A signal which is composed of real-valued data. For example: a series of temperature or pressure readings are typically recorded as integer or float type values.
Categorical Signal¶
A signal which is composed of discrete data. For example, a text (string), boolean type data or discrete numeric levels, such as the phase number of a process (1, 2, 3), the position of a switch (open, closed), steel type, alarm state (true, false), etc.
Sampling rate¶
The interval between two consecutive values in a signal is called its sampling interval. When measured in terms of the number of consecutive values per unit of time, this becomes sampling rate. The number of such values collected per second is defined as the sampling frequency.
Low sampling frequency¶
When a Signal has a sampling frequency of less than 100 mHz, i.e., 1 sample every 10 seconds or slower, it is considered to be of a low sampling frequency.
High sampling frequency¶
When a Signal has a sampling frequency of more than or equal to 1 Hz, i.e., 1 sample every second or faster, it is considered to be of a high sampling frequency.
Signal Tree¶
A Signal Tree is a non-overlapping organization of signals to form hierarchical categories of Signals. Those nodes of a Signal Tree that a Signal is associated with, are treated as the Signal's metadata. For example, if a Signal specifies two pieces of metadata, such as being part of a certain service/process and being of a certain measurement type, then that Signal becomes a part of two non-overlapping Signal Trees.
Signal Metadata¶
Any Signal Tree nodes that a Signal is associated with such as its location, measurement type, source, or other grouping are considered to be the metadata of that Signal. As a result, a Signal may have many pieces of metadata, each from a different Signal Tree.