Data Repository in Framework for Computation Pipelining: Model, Architecture, the Implementation Performance Estimates

In pipeline a single computation unit is deemed an estimator or standalone software component to evaluate statistics, estimates, test statistical hypothesis, etc. Template depends_on⟨⟩ is very restrictive and any declarative approach / language to define functional dependencies between estimates (features) will break the current automatic mechanism of BA’s pipeline composition. Whereas the CBD technologies have been studied over half a century the CBD standards are still far from state-of-art. The property “parametrized” we define as follows: an Estimator instance can be parametrized in common sense (i.e. with data structures to define algorithm parameters) as well as in general case (i.e. with model instances to interpret input data, parameters or with instance(s) of another Estimator(s) to implement superposition).