Analytics SDK provides APIs which does common functionality on analytics, which can be used by DAS services.

Common terminology and concepts are explained. APIs are listed <here>

Terminologies and Concepts:

1. Episode: An event that happens over a continuous period of time.
   For Eg If the episode is to find values to be 0.

2. What is EPS, Mag, and Duration which is passed as the output of each column.
To extract the second level of information from raw data, three new columns are created out of each input column.

• Eps: Is the value pass the rule of classification? Eg, if we are rule is to find the value is 0 or not, then this will be True if the value is 0 else False if the value is different than 0,

• Magnitude: If the rule(as explained in Eps) is satisfied, the magnitude can be used to quantify. Eg, if the rule is more than 35 or not. Magnitude can be the raw value where the rule is met. Both 50 and 80 pass the rule. The value of 50 and 80 can quantify the input data.

• Duration: This will give the duration of the episode. Eg, the rule is satisfied continuously for 10 minutes or 30 minutes. This will capture the continuous-time where the rule is met. Refer Episode section, the duration of the first episode is 3 minutes. and the duration of the second episode is 2 minutes.

3. Missing value(FMV):
If the value is missing/blank or a value of 0(configurable), we consider it as a missing value.

4. Sudden value change(SVC):
If the value changes more than usual, this event is captured here.
From the historical data, we find mean and standard deviation - which is treated as normal behavior. For the input we calculate the z-score based on historical stats, with the threshold on z-score, the input can be classified as sudden change or not.

5. Rapid value fluctuation.
If the value irregular rises and falls, this would capture this change.

6. Value out of range(VOR):
If the column has a few thresholds based on which it can be classified as the column is under the limit or how far it is off from the limit.
Eg: Oxygen saturation in humans:

7. Divergence path value(DPV):
If two values are expected to be in a similar range. If the difference goes beyond the normal behaviour based on history, this DPV will capture the unintended difference.