Introduction to Machine Learning

Preface:

Machine learning is a subset of computer science and artificial intelligence that touches on statistics and optimization, The field deals with the development of algorithms designed to allow the computer to learn from examples, and operates in a variety of computational tasks where traditional programming is not possible or very complex (a lot of hand-tuning or long list of rules).

I will present to you :

Why use Machine Learning, A little about the differences between the traditional approach and the Machine Learning approach.Some types of Machine learning System and Summary.

Let 's start with the definition of Machine Learning:

“A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T,as measured by P, improves with experience E.”

Tom Mitchell,1997.

Why use Machine Learning:

The main goal of learning the machine is to handle real-world data for a particular problem , when it is not possible to write a computer program for it, for example, a problem that a human expert can solve, but can not write the rules explicitly, or change over time and can not be written in advance.

the differences between the traditional approach and the Machine Learning approach:

The starting points for traditional approach and machine learning are quite similar, both aim to solve problems.

Let's look at the traditional approach:

first we would look at the problem then might notice about some pattern.
after that we write a detection algorithm for those pattern that we noticed.
and finally we test the program and maybe repeat the previous steps until it is good enough.

Suppose the problem is not trivial , the program can be a long list of complex rules and its can be very hard to maintain.

So the best and simplest solution to this problem is a machine learning approach:

a program based on Machine Learning automatically detecting the unusually frequent pattern and compared to regular examples.
it is give us a shorter program and easier to maintain.

The machine learning is great for:

Problems whose solution requires a long list of rules , one Machine Learning algorithm can often simplify code and perform better.

Complex problem for which there is no good solution at traditional approach, apparently a good Machine Learning tech can find one.

types of Machine learning System:

The supervised learning:

the training data contain the desired solution and feed to the algorithm .(e.g. Classification,predict a target value and so on…)

The unsupervised learning:

The data does not contain the solution inside,it like to learn without a lecturer.

Reinforcement learning:

select and perform actions, and get rewards in return.(e.g. touch the boiling water then burned)

Two main approach to generalization:

Instance-Based:

the system learns the examples by heart,then generalizes new cases using a similarity measure.

Model-Based:

model of examples ,then use that model to make predictions.

Summary:

A few words about the article, I tried to convey the introduction to machine learning in simple terms.
I wrote these words from summaries I prepared for myself , It was an opportunity to go over the material and share it with you.

In addition to those who want to delve deeper, I would recommend choosing a good book (Hands-On Machine Learning my favorite), write summaries and sharing your results with others.