The probabilistic feature provides functions and operators to process discrete and continuous probabilistic values in a data stream. Continuous probabilistic values are represented using Gaussian Mixtures.
To enable the probabilistic processing you have to include the probabilistic feature and issue the StandardProbabilistic transformation configurationĀ (#TRANSCFG StandardProbabilistic) in your Odysseus script.
For filtering probabilistic values you can use the same syntax that you already use for deterministic values. However, the result of the operators differ. In case of discrete probabilistic values the Select operator returns a tuple with a lower tuple existence probability.
Lets assume you have an attribute x and that attribute is 1.0 with probability 0.25, 2.0 with probability 0.25, and 3.0 with probability 0.5. The following Select operation will now filter the attribute value such that the resulting attribute value can only be instantiated to 2.0 and the resulting tuple existence is reduced to 0.25.
filter = SELECT({predicate = RelationalPredicate('x > 1.0 AND x < 3.0')}, input) |
The filtering of continuous probabilistic distributions is similar to the processing of discrete probabilistic values in the fact that it may reduce the tuple existence probability.
Lets assume you have a random variable x with mean 0.0 and Ļ2 1.0 the following query will set the tuple existence to the cumulative probability that this random variable will take a value between the upper and lower bound that is ~0.1586235826896239.
filter = SELECT({predicate = RelationalPredicate('x > 1.0 AND x < 4.0')}, input) |
The join with a predicate based on probabilistic discrete values uses the same syntax as for deterministic values. Although it looks similar the result is different in the sense that the Join operator performs a join of the input streams in each possible world and as such the operator may produce more tuple.
Select * From input1,input2 WHERE input1.x=input2.y; |
As you can see, the probabilistic processing is not limit to PQL. You can use the same CQL syntax you already used for deterministic values.
To access the tuple existence during processing you can use the ExistenceToPayload operator that copies the tuple existence to the payload where you can access them with the attribute name "meta_existence".
filter = ExistenceToPayload(SELECT({predicate = RelationalPredicate('x > 1.0 AND x < 4.0')}, probabilistic:data)) |
filter = SELECT({predicate = RelationalPredicate('x > 1.0 AND x < 4.0')}, probabilistic:data)