Social Network For Analytic Professionals

Which programming languages provide the fastest implementations, from specs to product?

Of course, it depends on the type of project. I'm talking about prototyping and testing scalable applications on datasets involving several gigabytes at least.

Text mining, keyword analytics: I would think scripting languages, Perl in particular, will help you prototype an application 10 times faster than traditional SQL (in terms of total number of hours of programming)
Marketing Databases: SAS can be very efficient if you use standard models (logistic regression) or if you have money (SAS Enterprise Miner)
Decision trees: maybe Java or C++? Althought I've found JMP to be very efficient / easy for small projects (much less than 1GB of data)
Hidden decision trees: SAS, Perl, SQL will all work very well, very fast
SVM: ?
Neural networks: ?
Multi-threading environments: ?
Where would R shine (compared to SAS, Perl, SQL, etc): ?

▶ Reply to This

Replies to This Discussion

Permalink Reply by Larry on November 12, 2009 at 10:49am: Some considerations...

Python - even thought its interpreter language its often hailed for its ease of development and speed

Go - this is a new language developed by Google. It supposed to offer the ease of Python and the speed of C,C++http://golang.org/; ▶ Reply to This

Permalink Reply by George J Bloor on November 16, 2009 at 4:13pm: DOORs, DODAF, powerpoint, EVMS...these are some of the tools used to capture specs; ▶ Reply to This

Permalink Reply by Larry on December 25, 2009 at 10:48am: C or C++ would be a strong consideration for SVM. I know there was some C++ source code on the Netflix Prize forums about applying some SVM techniques.; ▶ Reply to This