View Single Post
Old 04-10-2012, 02:14 PM
tcristo tcristo is offline
Join Date: Apr 2012
Posts: 23
Default Re: Assistance requested for "mature" learner

Originally Posted by avatarmh View Post

Part of my motivation in taking this course is to relearn coding to render tangible and useful my conceptual and mathematical understanding of some of the ML methods.

Unfortunately, it's been a while (decades) since I've done any coding to speak of (although I have found proven to myself that any procedural or declarative aspect of coding are still deeply embedded in my thinking, e.g., I have successfully coded some trivial stuff in Java recently). What I am certain I will have problems with is: I/O especially reading off of the web or data bases, i.e., anything but flat files.

Might anyone recommend to me:
1. Platforms (that are either free or fairly inexpensive)
2. Preferred languages (that are either free or fairly inexpensive)
3. Websites where I might find some code fragments to get me going and build off of.
4. Websites with examples and data.
5. What everyone else is doing that is considered optimal or near-optimal.

Anyone who is willing to take a more tutorial/collaborate interest in the request would be greatly appreciated. I would be happy to explore things I might offer you in return.


I suspect a lot of answers you get will be directly related to the domain you are interested in. Are you looking for the easiest way to code up and test some of your ideas (or support the homework in this class) or are you more interested in developing something for use in a production environment that has some time based constraints?

I personally use whatever language and platform makes the most sense at the time and within the constraints given. I personally wouldn't use Excel\VBA to code up a client\server based credit determination system for a large bank. However, for a small community bank that has a single loan officer, I very well may do exactly that.

Tell us a little more about your interests and what you want to accomplish and I am sure you will get some valid input.

BTW: Almost every programming language has greatly simplified the task of performing I\O nowadays. I can read and parse diverse file types such as CVS and XML files using classes that are provided as a part of the compilers. If you can understand the theory of ML and code the theory into your own program, I seriously doubt you will have any problem loading the data to test and train it.
Reply With Quote