Showing posts from July, 2013

Lorem Ipsum 2 - bits of Hyderabad

These Lorem Ipsum posts are quite useful. They serve to plug in the lacuna that this blog suffers at times, while being faithful to the concept of a traditional log-book. Also these are periods of when I am unable to come up with anything better to write about, and am thus forced to conjure drivel in the form of a blog post.

I've landed in the fine city of Hyderabad, and have started my internship in an equally fine company. The work is brilliant, as are the co workers, and I'm staying at this neat place in one of the more recently developed regions of the historic city. So yes, I am far from the old city with all its Nizamesque culture and the quintessential biryani - and closer to shiny glass buildings and buttoned up code crunchers who take to the roads in hundreds of thousands every morning on a weekday.

The weather is absolutely fanstastic. Apparently after the torrid summer, which I'm thankful to have just missed, the weather remains like this till the onset of the …

Machine Learning with R - 1 : Decision Trees using the party package

Fueled by my new found programming love in R, and my general interest in machine learning, I've been looking at various online resources to integrate the two ie, use R to implement popular machine learning and data mining algorithms. Of course, machine learning is a fairly well established field in computer science today, and it has engendered several powerful algorithms - from Bayesian classifiers, to decision trees, to neural networks and the quintessential support vector machines. In this series of blog posts, I'll try to deal with some of these popular algorithms using R. You can access just this series of blog posts by clicking on the "Machine Learning with R" category in the tag cloud on the right.
Thanks to a brilliant open-source community, R has a vibrant developer group, as a result of which it has a pretty cool arsenal of packages that makes our job simpler. Our focus in this series is to utilise those packages and get some data crunching done.
This is th…