Hello Monkeyloggers. I have been away for a week doing unfortunate but necessary things, and I haven’t forgotten about the MonkeyMiners. This morning I published an article that will provide some insights to new modellers explaining that it is not always best to simply trust the predictive models; that with a little bit of your own insight results can be improved by many times. If you are curious, log into the article: Data Mining 101: Choosing Yes or No, Or a Powerful Maybe.

Work on the book proceeds, but we need to add examples, and that is slowing things down a bit. Should we add the examples after each node is introduced for a complete reference to each node and how they work, or put the nodes at the start of the book and add the examples at the end, so the start of the book remains a high level reference. I am struggling with this choice right now. If anyone out there has a preference, let us know.

We are also in the process of discussing what we can do that would make us different from other data mining and predictive analytic resources out there. We have been discussing specializing in just one tool, and perhaps creating a guide for people who are starting out who want to do predictive data mining on the cuff. I am leaning toward teaching how to data mine without investing a lot of money. Your ideas would be appreciated.

One last thought. I have been looking at Facebook groups to get an idea of what’s out there for data miners getting started. What I found made me feel small and insignificant. There are a lot of smart people out there doing a lot of amazing things, that are well beyond a knuckle dragger like me. Maybe there should be a group for beginners, one that we can start. Anyone else out there have thoughts on that, we’d like to heard from you.

Goodnight everyone.

How to interpret models
Tagged on: