Machine Learning Guide (Hacker Style)
Machine Learning Guide (Hacker Style)
The goal of this guide is to experience the power and value of Machine Learning, while gaining a good understanding of the concepts involved, in a quick and efficient manner.
At the moment, there is a huge number of online courses that claim to teach you Data Science and Machine Learning. Andrews course, for example, it’s hugely respected, and everyone attempts it, but most everyone gives up. The content is fantastic, but by the 5th video you start wanting to implement the concepts you’ve learnt and know that you have another [how many more videos are there?] more videos to watch before you can.
My goal was to gain a good grasp on Machine Learning as quickly as possible. You will not go deep into any one topic, but you’ll get a great overview by actually doing real things! Listed below are my thoughts and approach. I hope this helps you, as it helped me. Good Luck!
Generally, there are three things you need to study or learn something quickly:
- Great mentors
- The passion and dedication to carry through to the end
- Great resources that also suit your style of learning
In this guide, you will be responsible for the first two points, but regarding resources, this is what I highly recommend:
- Machine Learning A to Z - This uses Python and R, and can be found on Udemy (https://www.udemy.com/machinelearning/ )
This course is by far the best you can get online. It quickly shows a problem set and takes you through all the algorithms completely hands on from the first topic. I found this course after googling a lot and I must say that this is the best course online for a quick hands on and overview of ML.
- Intro to ML - This can be found on Udacity ( https://www.udacity.com/course/intro-to-machine-learning--ud120 )
Please do the mini-project for each of the topics. IT’S A FREE COURSE.This course is taught by the cofounder of audacity (Sebastian Thrun) who works on google self driving cars and google brain.
- Datascience A to Z by Kiril - This can be found on Udemy ( https://www.udemy.com/datascience/ )
- Udacity’s Nanodegree for Machine Learning ( https://udacity.com/course/machine-learning-engineer-nanodegree--nd009/? )
Start solving Kaggle problems and start from the basic level to level 3 kind of questions I’ll share 15 links which would be more than enough for you to get comfortable with Machine Learning.
I've started with Titanic problem and I've tried everything under the sun and I learn more by submitting each solution. I will write a separate blog on this. I’ve made more than 37 submissions for the same problem and have gone from 5500 ranks to within top 300. I’ll mention all the approaches later.
Currently, I'm planning to get my rank within 50!!
Also, I’ve tried everything from ensembles, stacking and neural network for the same problem :)
I haven’t done Fast.ai but I’ll go through it and let you know my feedback :)