[PYTHON] You will be an engineer in 100 days ――Day 76 ――Programming ――About machine learning

Click here until yesterday

This time we are talking about machine learning.

About machine learning

Recently, due to various reasons, I have been closed or unable to go to work. How about studying at this opportunity?

About machine learning without using mathematical formulas I would like to explain with contents that even elementary school students can understand.

Click here for the explanation video

What is machine learning?

First of all, what is machine learning? From that point.

Machine learning lets a computer learn data It is an attempt to make predictions for unknown data.

Learn to remember patterns hidden in large amounts of data The rule for judging unknown data is called a model. Making this model is the point of machine learning.

Machine learning and AI

Next, I would like to talk about the relationship between words and artificial intelligence these days.

Artificial intelligence is called ʻAI in English. Most of ʻAI is based on machine learning. And machine learning is called machine learning in English.

Deep learning is one of the machine learning methods and is called deep learning in English.

The relationship between words is as shown in the figure.

Where is machine learning used?

When it comes to where machine learning is used, it is now widely used in various fields.

The main fields are image discrimination and object detection performed for autonomous driving, classification of people and animals, and discrimination of handwritten characters.

It is also widely used for data related to human behavior. Whether to join or withdraw from the shopping site Machine learning is often used to determine what products this person recommends.

You can also generate various things from what you have learned. You can learn the style of the image and create an image that looks like it. Machine learning is also used to translate English sentences into Japanese to create songs that look like that.

What you can do with machine learning

There are three main things you can do with machine learning Regression Category Clustering is.

Each Regression: Predict numerical values Category: Predict categories Clustering: Divide into good feelings Can be said.

How to learn

There are three main learning methods Supervised learning Unsupervised learning Reinforcement learning is.

Supervised learning is a learning method that has an answer and lets you learn according to that answer. Unsupervised learning is a learning method for those who do not know the answer. Reinforcement learning is a learning method that maximizes the value of a game based on some value, such as winning a game.

Actual data

Data is performed using tabular data with vertical and horizontal directions. Normally, there is more than one type of data (column). In most cases, there are multiple.

This type of data in the column direction is also called a variable in machine learning.

Supervised learning

About supervised learning

Consider going for a price because of the size of the room. First, prepare the data (room size and price).

Think of drawing a nice line that fits your data.

Isn't it possible to draw a line like the one on the right?

This line is called the learning model (discrimination rule).

Let's make a prediction from the data using the created prediction model.

Looking at the part of the line that applies to it from the room size data The price will come out. This is the predicted value.

Prediction only gives a value close to the answer, and it is quite difficult to guess exactly.

Also, there is more than one data, so while looking at various data You need to think of a way to draw a nice line.

This time, I took the example of relying on numerical values, but this is called regression. Those who rely on non-numerical values such as male and female are called discrimination.

Unsupervised learning

Next is unsupervised learning.

Unsupervised learning is a learning method that divides into good feelings like clustering.

There is no answer, so it is up to humans to decide whether the predicted result is correct.

About deep learning

About deep learning.

It is a learning method using a learning model called neural net. With a learning model that imitates the structure of the human brain, you can set three layers: input layer, hidden layer, and output layer.

Of these, those with two or more hidden layers are also called deep neural networks (DNN). The learning method using the DNN is deep learning.

It's a mess, but what's great about deep learning is It means that the machine has exceeded human precision.

Especially in the worldwide image recognition contest A model that exceeds human accuracy has been created.

As a result, it has come to make more accurate discrimination instead of humans. It has become active in various fields such as translating English sentences into Japanese.

What you need for machine learning

For those who want to start machine learning for the first time, prepare a PC first.

Next, let's install a programming language called Python.

Finally, prepare tabular data to be used for machine learning.

Well, if there is someone who wants to do machine learning but doesn't understand programming We have prepared a course, so please refer to it! !! !!

reference: Python programming course

Summary

It's a hyper-rough explanation about machine learning.

Machine learning itself has been adopted by every company in this age. It has become an indispensable tool for services.

Therefore, it is becoming difficult to do without knowing it.

If it's just a concept, it's not too difficult. If you try to pursue it in detail, there is no end to it.

First of all, acquire it as a rough knowledge How to realize it when it comes to doing business etc. I think you should learn.

24 days until you become an engineer

Author information

Otsu py's HP: http://www.otupy.net/

Youtube： https://www.youtube.com/channel/UCaT7xpeq8n1G_HcJKKSOXMw

Twitter： https://twitter.com/otupython