I recently had an opportunity to work on a supervised machine learning problem where the target variables are real and multi valued. An initial search of the machine learning libraries revealed that popular Open source ML libraries have little support for the same.

Here's a quick intro in the topic, and (in a later post), a dive into some of the libraries from the JVM ecosystem.

Open source in the software industry is a well known phenomena that needs no mention. In the Analytics industry, while the use of Open source software (OSS) is extremely common, there is certainly a need for OSS platforms that will enable Analytics service providers to collaborate rather than compete.

Sharing an article that I wrote earlier, from the Bridgei2i Blog

Gorilla-repl is a toolset from the Clojure stable, which enables us to interact, visualize and share Clojure code as a worksheet. For those familiar with IPython notebooks (now called Jupyter), Gorilla-repl is the Clojure equivalent of the same. I find that it is an excellent environment for sharing and discussing data science.

Using Spark's Machine Learning classifiers in Clojure

The Sparkling project enables Clojure developers to use Apache Spark (which is primarily developed in the Scala langauge) for big data processing and machine learning.

Html Templating in Clojure.

Clojure's Lisp root (where code-is-data-is-code) lends itself nicely to generating HTML. Enlive is a promising library capable of templating, and here I present a tutorial for Enlive.