Group Message Board › Talk last week about statistics
This message board is read-only.
Fort Lee, NJ
I have posted a press release in the FILES menu for a meeting in California of the ACM, a serious computer science group that does a lot of work in machine learning. As I suggested the machine learning people are very interested in Bayesian statistics. In NY the Machine Learning meetup is having a meeting on this topic this week. The press release does a great job of describing Bayesian approaches.
Getting back to Lynda's question about why do we do still use the 100 year old approaches, e.g. frequentist statistics. Don't think of it as a tool because then you will get the response, "when all you have is a hammer, then everything is a nail." It is a theory, like Newtonian physics that works in most applications. But I would think of it more as a repetoiry, a classic that can be adapted and interpreted for the scientist, customer and analytic professional.
Applied statistics is a bit of a craft and collaboration, like theater. For example, Anything Goes is on Broadway again. It works because it serves the purpose of entertaining tourists and groups, reinterpreting classic songs for theater regulars, showcasing stars, and offering new talent casting opportunities in a known entity. Of course there have been improvements since it was written about 80 years ago. But other musicals, and statistics, have faded into obscurity.