Machine Learning

It has become one of the most sought-after skill by enterprises. You may be wondering why? So let me answer this first -

It is when we stop programming and start TEACHING computers to learn from data and computers starts doing what you want it to do! Welcome aboard, you are using Machine Learning.

So everyone told you that the next big thing in IT industry is machine learning or better known as ML.

What does Machine Learning do?

It is good to know what is the purpose and business model behind a technology (always remember no one creates or manufactures a product that doesn't have business value). Machine Learning makes apps and devices smarter!

We have already experienced this?

Do you ever wonder how an online shopping app / website suggests you product or services or why your credit card was blocked when you swiped it at a particular POS?

The secret is out and next time when you are suggested it is not done by any homo sapiens sitting behind desk and monitoring your purchases instead it is done by a system which is using machine learning for product recommendation based on what you've purchased. When your credit card is swiped, machine learning compares the transaction to a database of transactions and helps detect fraud. When your robot vacuum cleaner vacuums a room, machine learning helps it decide whether the job is done.

All major cloud computing providers are offering ML infrastructure which enterprise can use for predicting consumer behavior, customer sentiments, detect frauds etc. and create targeted marketing strategies and fine tune their business plans to derive maximum benefits.

Is machine learning is an apt skill for you?

A million dollar question! I get lots of mails, phone calls from many graduates, who are just out of their college, or IT professionals working with IT firms ask - Shall I do a course or obtain certificate in Machine Learning? Will learning ML help me in advancing my career or get me a job?

Let me try and answer this question by asking you few questions

• Are you fond of statistics and have fundamental understanding of statistics?

This is a two part question and answering this question will help you in deciding whether you should go for it or not. Numbers and statistics may not excite many people but if you want to make career in ML you have live, eat and sleep with statistical algorithm. A passion to understand statistical models. Some of the most important concepts relating to statistics that you must know are - types of data (numerical and non-numerical), mean mode and median. Understanding of Bayesian statistics. If you love statistics and find your passion to use it over the data Machine Learning is a good choice.

• Do you get easily bored when surrounded by huge volume of data?

If yes, I am sorry you will not have enough motivation to stay in this field. ML is all about working with humongous amount of data (however, machines work with the data but you need to be around for gathering and cleansing.

• Have you on worked on ETL (Extract Transform Load) tools or worked on ETL projects?

It will be great if you have worked but will not make much of difference if you haven't. ETL involves EXTRACTION of data (i.e. connecting to different data sources (db, flat files, web services, stream ) and extracting (pulling the data from these sources)) TRANSFORMING (bringing the data to a central repository and making it pass through a pipe line to change data type, cleansing, combining and finally LOADING the data into a central repository (like data warehouse) and perform analysis using different tools.

How do I start learning Machine Learning? What is my learning Path?

As soon as possible. The approach we will follow is - I will share step by step video tutorials that you may follow and want to do it yourself. Beside this during the course of this learning adventure I will explain and help you understand the different skills and knowledge you must posses if you aspire to make career in Machine Learning. One important thing to keep in mind, I will be using Microsoft Azure cloud in these videos. If you want to learn along you can obtain Azure Free Trial subscription by visiting this link. All you need is uan nused credit card (i.e. you have not used this card previously for obtaining azure subscription). The Azure free trial subscription will give you approx $200 worth credit for 30 days (and 1 year access to some services) and Microsoft will not charge you or will not renew your subscription without your consent.