Big Picture of Machine Learning
In this article series, I’ll give you a step by step basic theories and hands-on experience practical articles to get the basic knowledge in the machine learning path.
First of all, let’s get a brief introduction to machine learning.
What is Machine learning?
This is the very first question that you all have who are willing to learn machine learning.
Actually, What is machine learning? Answering that question Simply, Machine Learning is “The field of study that gives computers the ability to learn without being explicitly programmed”.This the very easiest and understandable answer to all the very beginners in the machine learning path. Remember that, This definition is a very informal one and it is only for your easy understand. There is a much more very formal definition of machine learning and I’ll give it later in this article.
Why do We need Machine Learning?
The next burning questions are “Why do we need Machine Learning?” and “Is it essential for day to day life?”. The answer for the first one is “because we want to complete our day to day works easily and spending less time, for that we want to automate certain processes with fewer errors (Machines do fewer errors) using a computer”. Now you’ll get another question that “We have automated certain computer processes for the last 60,50 years, why is this machine learning?”. So, think that there many use cases that we can’t program or automated by hand. Because those use cases are that complicated to think and program by hand. I’ll give examples of this later in this article.
Answering my second question which is “Is it essential for day to day life?” , Simply, it is yes. Because we know in the 21st century which is technically known as “Data Era”. So we can’t live without data and without machine learning, we can’t give a value for data.
So let’s move to the most interesting part of this article.
Use cases of Machine Learning?
When we consider the use cases of machine learning we have to refer in several categories. Firstly,
1. Data Mining
In modern advertising, the valuable thing is data, Due to that reason they use machine learning to mine data to find hidden patterns and trends. For example, web click data is one technique used by the modern marketing world to find the likes and dislikes of the customer. According to that, they can do their marketing thing very efficiently and effectively.
As well as in the medical stream, they use Data mining to diagnose diseases by using past medical records and find patterns and trends regarding diseases. For example, to predict a given the patient’s tumor(harmful) is malignant or benign(not harmful).
There are so many use cases regarding data mining in various streams.
Secondly,
2. Applications can’t program by hand.
There many use cases in the world some times we can’t imagine how to program it to work automatically. Because those use cases are that complicated to think. If you think about that we have to consider lots of aspects in that use case. Usually, it is not practical in today's world. So here again machine learning helps us to solve that issue,
for Example Autonomous vehicles, helicopters, Handwriting recognitions, most natural leaning processing, and computer vision.
Thirdly,
3.Self-Customizing Programs
There are some use cases that programs can think themself perform several tasks. Such that Email spam filtering, Google news, Netflix, Amazon product recommendations and etc. In this case, they also use some machine learning techniques to give the ability to the machine to think themselves and perform some tasks.
So Finally, you have some better basic idea of machine learning as a beginner. So let’s move to a formal definition of machine learning given by Tom Mitchel 1998.
Machine Learning is,
A computer program is said to learn from experience “E” with respect to some class of tasks “T” and performance measure “P”, if its performance at tasks in T, as measured by P, improves with experience E.
To explain this think about an Email spam filter System,
- There, first, we classifying emails as spam or not (Task T).
- Then we give these categorized emails to the computer and then the computer learns about the spam mails and not spams (Experience “E”).
- Next, we give a new mail set to the computer and watched “how many emails correctly categorized as spam or not (measure performance “P”).
So, finally, I gave you the formal simple definition of machine learning and give you the big picture of machine learning. I hope you got a better introduction to machine learning in this articles as a beginner.
