If you are wondering what the heck is Machine Learning (ML), how do I learn it, what skills do I need to start learning it, then you have come to the right place. This blog is for you! In this blog, we are going to guide you through the basics of ML and how you can master them too. After reading this blog, you should have a fairly good understanding of the field, how and where you can use it.
This blog is divided into following sections
What is ML
Why should I care about it
Where do I use it
How do I begin learning ML
What next
1. So, what exactly is Machine Learning?
First, let's talk about Artificial Intelligence.
Artificial Intelligence (AI) is the broader concept of machines being able to carry out tasks in a way that could be considered as smart. The machines not necessarily learn from the data but may exhibit intelligence in performing certain tasks that mimic the characteristic of human intelligence. AI consists of lot of sub fields such as Machine Learning, Natural Language Processing, Knowledge Representation, etc.
Now, coming back to Machine Learning. It is a subset of AI that aims at making systems learn automatically from the data provided and improve their learnings over time without being explicitly programmed.
ML can be divided into 2 major components:
Supervised Learning - where your training data consists of output labels as well, which you want to predict. For instance, predicting whether a given email is SPAM or not, given sample emails with the labels whether they are SPAM or not, falls within Supervised learning.
Unsupervised Learning - where the training data does not contain the output labels. The objective is to group the similar data together instead of predicting any specific value. Clustering, Dimensionality Reduction and Anomaly Detection are some of the Unsupervised Learning techniques. For instance, grouping the customers based on their purchasing pattern.
Some of the best Machine Learning resources for beginners are:
Andrew Ng’s Machine Learning course on Coursera or Stanford CS229.
Machine Learning Yearning. You can download individual chapters for FREE.
An Introduction to Statistical Learning by Gareth James et al, available FREE online.
2. Okay, I think I understand the meaning of ML but why is it so important today?
Looking at the past trends, the world data is expected to double every two years with the cost of data storage declining at roughly the same rate. In order to extract the information from this exponentially increasing data, the need for better Machine Learning techniques is at an all-time high.
Machine Learning is perhaps one of the hottest fields in the world right now, not just in Silicon Valley. Every industry, every company is using Artificial Intelligence and Machine Learning in one way or another. For any company wanting to gauge the reach of their business and products, using Machine Learning techniques is the only scalable approach in the long term.
It is not surprising that the demand for Data Scientists and Machine Learning engineers is more than ever before. And, I believe that it is extremely crucial to keep yourself up to date on the latest research and developments in this field. Be it automation, personalization, cognitive services or predictive analytics, Machine Learning is going to be the center point of all such fields.
3. Awesome! ML, indeed, sounds very exciting. I didn't realize it would be this important. Looks like everyone is using ML in some way or the other.
Yeah, you're right. You can see the applications of Machine Learning probably everywhere around you, for instance in gadgets such as Amazon Echo and Google Home, personalized ads shown on Facebook, Video recommendations on Netflix and Youtube, Product recommendations on e-commerce platforms such as Amazon and Flipkart. The major use cases of Machine Learning lies in analyzing Big Data, creating Predictive Models and Deep Insights.
4. Wow! So, not only the big companies but even small to mid companies also make use of ML. How can I get started?
Here is the interesting part you have been waiting for! A lot of people think that they can just use a ML library and with publicly available data, they can directly start building ML models. While this is true, but this is the wrong approach.
Even if you don't want to be a ML expert, you still need to learn some basic fundamentals of Mathematics and Statistics to apply ML in your case. For starters, you should definitely have a good understanding of probability and statistics, linear algebra, distributions, mean, variance, few statistical measures such as p-value, f-value, mean squared error etc. This would greatly help you understand and analyze your ML model.
Once you have some decent understanding of these concepts, you can jump into the coding part. Remember, it is an iterative process and you would have to keep switching between mathematical fundamentals and programming parts to appreciate how ML algorithms work.
ML algorithms and libraries have been implemented in a lot of programming languages such as Python, Java, C++, Matlab etc. For beginners, I believe Python is the best language. There are few reasons behind using Python:
Because it is the most human readable language and easy to learn. Users can focus on machine learning rather than exhausting syntax and data structure errors.
Your ML model revolves around the data. Data is the key in any ML workflow. And more often than not, it would be raw, unstructured, garbled, huge, incomplete etc. Here, Python packages can come to the rescue. Python offers a huge number of packages which you can use with various types of data. For instance, numpy, pandas, scikit, matplotlib, nltk, scipy. They are all Free to use under GNU license.
Best courses for Python beginners:
Python Fundamentals This is a great course to learn Python. It starts from ground zero and covers all the major topics.
Complete Python Bootcamp: Go from zero to hero in Python 3 This is one of the most popular course to learn Python on Udemy and more than 250,000 students have enrolled in it.
Intro to Python for Data Science This is an in-depth and comprehensive course and covers everything you need to know.
5. Awesome! Can't way to get my hands dirty and dive into ML. What now?
Once, you have Python installed on your system and know basics of programming in the language, it's time for you to start your ML code. Take any framework of your choice, TensorFlow, Keras, PyTorch, Scikit-Learn etc. Understand their documentation and start coding from their Getting Started page.
So, that's all for this article. Hope you found it useful and resourceful.
We have also written a book on Machine Learning "Cracking The Machine Learning Interview" which covers all the major topics of ML and is very useful for beginners in the field. You can check it out on Amazon.
