Computability

An understanding of computability can help data scientists recognize the limits of algorithms. The feasibility of algorithms is an important consideration before embarking on substantial data science initiatives, before significant time and investment have been allocated toward presumed solutions. It could be that a problem is not solvable or that it may only be solvable after an extraordinary outlay of funds. It would be best if, in advance, the decision to proceed was based on the proven logic of computability . Data science is concerned with empirical proof, evidence based on observed conditions from sets of data. However, observation is not always immediately possible because computations can take significant time. For such cases, mathematical logic expedites analysis of the relative performance of algorithms. Empirical data is that which is encountered outside a model. Such data may be considered factual and may be recorded from the outside world. In contrast, logical data is internal to models, as it may be assumed or induced logically. The overlap of these concepts with respect to machine states is the essence of computability .