readme.MD

Machine learning - 2016

Course on machine learning organized by [Yandex Data School] (https://yandexdataschool.com/) at Imperial College.

Updates on the course

News are published on gitter chat. Write there about any appearing problems. Gitter uses github accounts (so you don't need to register).

Download first pack of lectures

Initial preparations (getting access to cloud)

1. register at github
2. register at kaggle
3. login to everware with your github account. Put docker:arogozhnikov/mlatimperial2016 in the field for url. Everware will start a personal container (with copy of this repository)
4. (at everware) in the folder MLatImperial2016 click on notebook 0 - Initial notebook. Run it (Cell > Run All) from the beginning up to the end. It should generate a basic submission and provide a link to download it.
5. download a submission and upload it to kaggle challenge.

To get an access to the cloud again, just go to everware - it will keep all results of your work. Please don't logout from everware - this will destroy container with all your results.

To dive faster into data analysis it is suggested to study the essentials of python and its scientific libraries using recommended tutorials. On the first seminar we will introduce them to you but to complete the excercises you will need to study most of them yourself. You can practice your skills by creating test ipython notebook on everware or install python on your computer.