Manipulating Predictive Analytics
In my blogs entries, I tend to mention FACTS.
A fact is something that I have actually accomplished. At times, people think that I am being theoretical; however, I tend to tell people when I am being theoretical and when it is something that I have actually accomplished.
For example, here are a few “Zimm Facts”:
- I have implemented, administered, and/or enhanced a number of predictive analytic engines for a number of different industries.
- I have played with the data requirements necessary to run a predictive analytic engine. In some cases, I failed. In more cases, I had results that I still do not accept. (e.g. 24:1 return. What the? Who the? And yes, I quadrupled checked the numbers.)
- Predictive analytic engines work better on limited data attributes and through the incorporation of model scores.
- Predictive analytic engines work best when they incorporate both Big Data and Cloud Data.
So here is where I am going to get a tad bit theoretical for all of you.
Theory: Predicative analytical engines can be manipulated with the right Cloud Data.
Let me walk you through this one.
Since predictive analytic engines work better on limited data attributes and through the incorporation of model scores, and strong predictive analytic engines work best when they incorporate both big data and cloud data, the predictive analytical engine can be manipulated with the right cloud data.
If you want proof, look at this article: “Twitter Mood Predicts the Stock Market’. This is only one of a number of examples that I could give.
But as I know how the predictive analytic engines work, my question is: “Did Twitter predict the stock market, or was the stock market predictive engines manipulated by twitter?”
If my logic is sound, which I am 90% certain it is, the following theory should hold true:
If a company creates the right cloud data (via social media), that cloud data can have a direct impact on existing (internal and external) predictive analytic engines through the data and models that feed those predictive engines.
Think about it for a moment, and then think “How cool is that?!” Yes, I do geek about this stuff.
Now I just need to find someone who will let me play with their social media program and test my theory.