A starting approach to build a data science, or a cognitive model, is to work from a business process model. Ultimately any improvement changes a way of doing things. Making it faster, better, cheaper. Even a very simple, even sketch of a model can remind us what we are doing, what data we need, and how we can measure the results.
Business process models (BPM) have been doing this for some time for more complex processes. And during an earlier, cognitive era, we directly used rules and process flow to model the reasoning parts. In machine learning, there are methods like decision trees, which more directly model a portion of a process. Decisions are by their nature cognitive.
Yet still have not seen enough integration of BPM and cognitive. Especially in a way they could be used to sell and deliver starter proof of concepts. To clients or management. And to experiment with alternatives in their implementation and value.
I was told that IBM was experimenting with the idea, but had not seen much input. Then came upon the article here
Not yet convinced that this is simple enough to construct the kind of easy templates for Process+Cognitive+Analytics I would like to see done. But it does appear to be a good start. Digging deeper. I also like the mention of Sense-Respond-Learn which takes it to stronger elements of cognitive.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment