In year 1959, Arthur Samuel used the term Machine Learning(in short ML) for the first time.
Lets see the formal definition of Machine Learning:
According to Wikipedia–
“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.”
But understanding above statement is bit difficult. So lets try to go deeper to understand machine learning better.
Where Machine Learning is being Used?
As in this Tech world Machine Learning can be used anywhere. As shown below:
For simple example, if you have ever done Andrew Ng ML course, you’ll come to know you can predict house prices, according to some parameters like: where house is located, number of bed rooms, area and so on.
It is very hard for us to do the same thing if given a huge number of data sets. But a Machine Leaning model is capable of doing such things in a second. So in such case I guess you should ask a ML model for help.
What basic things you need for Machine Learning?
Collecting training datasets: First, you need a training data set on which you should train your machine learning model. A data set can be taken from anywhere it depends on you. If you need any kind of dataset please let us know in the comment section. Otherwise, you can collect your datasets manually (with fewer errors) or it can also be done automatically with some python scripting.
Create a ML Model: After collecting datasets you need to create a machine learning model to make your predictions according to what you have given input to that model.
Train ML Model: You need to have a very varied training datasets to get predictions more closer to actual results.
Machine Learning Algorithms: You may have many tasks related to machine learning or just one. There are many algorithms to help you in all cases but you should choose the best machine learning algorithms that suit your requirement best. The model you are going to build should have good speed and it should produce results accurately. And speed and accuracy of the model you have build depend on what algorithm you are using.
What you need to Learn?
As we all know Machine Learning is very closely related to Linear Algebra, Statistics and Python.
Linear Algebra: Before going for Machine Learning you must have basic understanding of Linear Algebra. But you don’t need deep knowledge of Linear Algebra because in Python libraries already many of the math algorithms are included. Along with this if you want to develop your own machine learning models you need to learn linear algebra from scratch.
Statistics: Any machine learning model you see is build upon data. While working in Machine Learning domain you’ll be searching for best dataset for your model. And working with data needs knowledge of Statistics. So you should clear your concepts about Probability Distributions, Regression, etc.
Python: Python is a higher level, object-oriented programming language, which is most popular among machine learning learners. Its popularity is just because it has many useful built-in libraries for machine learning. Python community is huge and you can get help easily.
That’s all about this blog. Hope you’ve now better understanding of “How to Get Started with Machine Learning?“
For reference watch below video: