BorutaPy – an all relevant feature selection method

danielhomola Blog 8 Comments

TL,DR: There’s a pretty clever all-relevant feature selection method, which was conceived by Witold R. Rudnicki and developed by Miron B. Kursa at the ICM UW. Here is its website. While working on my PhD project I read their paper, really liked the method, but didn’t quite like how slow it was. It’s based on R’s Random Forest implementation which runs …

danielhomolaBorutaPy – an all relevant feature selection method

Sending emails from Python through a Gmail account

danielhomola Blog Leave a Comment

In my current research I’m building a new research tool for the integration and visualisation of genomic data. The application starts with a file-upload form, then runs a pretty complex pipeline on our server and cluster that can take hours to complete. So once it’s finished I need to notify the users about any errors, warnings, send them the result files and …

danielhomolaSending emails from Python through a Gmail account

Data science and machine learning podcasts

danielhomola Blog Leave a Comment

This is just a quick one.. Generally it’s quite hard to be up-to-date with all the amazing stuff that’s happening in data science and ML especially if you’re doing a job that’s not a 100% related to it.. I used to just read random articles and follow links in blog-posts, but then I found a few resources that really help …

danielhomolaData science and machine learning podcasts

New website

danielhomola Blog 2 Comments

Hurray! I redesigned my website yet again.. I think for the 3rd time since its inception. It was about time as I haven’t really changed or updated it in the last 2 or 3 years which is a shame.. But things are going to be very different from now on. Yet another new year’s resolution.. I’m planning to write at least …

danielhomolaNew website