To my best understanding, the science revolution as envisaged about 400 ears ago can;t be stretched to cover to big data analytics, first and most because it never had to cope with the kind of issues that are being hurled on us now by big data, such as dat originated in machines and dynamic reality. I propose to redefine the scientific fundamentals, I have a list... anyhow the common denominator and the overwhelming need is for making new hypotheses, in addition to verifying them.
I agree that a new scientific inquiry is needed with unlimited and dynamic access to data being possible. The NIH Big Data to Knowledge is a good start but it seems to be applying yesterdays approach to today's opportunity. There is a urgent need to find a new paradigm (of evaluation, rewards. incentives) for open science that is now possible bringing with it a quantum jump through translational research in all fields.
Allow me to start from the middle... by checking new terms that entered to the world of analytics. For example - PATTERNS - the collections of phenomena that share common causes. In some models (including mine - GT data mining), identifying the patterns is mandatory for analytics. I strongly believe that finding the patterns is a necessary first step in analytics and the new Science revolution should include logical rules for searching and validating new hypotheses regarding patterns. I mean all patterns including exceptions, rare, newly emerged trend, hidden and concealed ones, or in other words - patterns that we have no prior knowledge about!
Didn't Darwin dig for patterns to classify species? Isn't physics about finding patterns to construct models of the world? Isn't any experiment an experiment to discover patterns and then try to explain them or work with them to gain knowledge? What's new about patterns?
Nicolai, true that patterns are an old notion, but in the old time they were observed through induction, while now (due to big-data complexity and dynamics) they need to be discovered in first step, through deduction.
Now days when an expert discovers a pattern, he does not have necessarily to know what are the interrelationships, all he needs to know is the pattern scope and characteristics. For example: discovery of yield and quality patterns in production, or a new weather pattern that drill-down analysis later finds to be caused by East ocean warming.
- Badges
- Science topic