[PYTHON] 5 masterpieces of data science that you can learn for free [Japanese books are not scary if you can speak English even if they are expensive]
Let's study by ourselves at this time
Perhaps there are many generous organizations overseas, there are so many specialized books that can be read for free.
There are other summary articles of this type, I would like to display the translated Japanese books and the original books side by side. Introduced only in the field of data science.
That expensive masterpiece may actually be free if it is the original.
(As of May 2020)
1st book
Japanese title: Basics of statistical learning (15,000 yen)
This is the so-called "castella book".
――In Japan, it was translated in 2014, but the original was published in 2001 and about 20 years ago. From the content, it must have been a novel book that systematically studied theory at that time. It is exactly an "algorithm / theory dictionary". It feels old in the era, but the basics are always important. If you are not strong in numbers, you will never be able to read it. .. ..
Original: The Elements of Statistical Learning
Abbreviation: ELS
The thickness of the book, the color of the cover, I really want to go to Castella, Nagasaki. The contents are so heavy that the stomach seems to lean.
2nd book
Japanese title: Pattern recognition and machine learning (top / bottom) (16,000 yen)
Known as "PRML"
――When you start machine learning, this book is as recommended as "Hajipata (first pattern recognition)". I hear people around me calling it "bishop books".
――It is not a book that learns algorithms evenly like the first book, but it theoretically introduces pattern recognition and the algorithms around it.
-ctgk explains the problem in notebook format, [Darkside communication group provides supplementary material](https://herumi.github.io/ prml /) You can see that it is a masterpiece from the fact that everyone is assisting us in making it.
(Both "Hajipata" and "Bishop Book" kill beginners, so it's hard if there is no one to teach.)
Original: Pattern recognition and machine learning
3rd book
Japanese title: Introduction to statistical learning with R (7,000 yen)
Known as "ISL"
――Compared to the above two books, I feel that the name recognition is that much, but it is a well-known book. --The concept is that if "Basics of Statistical Learning: ESL" is a systematic book written for those who understand mathematics, then this book is more written for beginners. That is. ――It is true that the content comes from the basics of important ideas in data science.
However, it is difficult for someone who does not know anything. I think it's better to have knowledge about probability distribution and regression analysis in advance so that you don't have to worry about it.
The sentences are long. I think it's done carefully to make it easier to understand, but sometimes the story doesn't come in.
Original: An Introduction to Statistical Learning with Applications in R
4th book
Japanese title: Deep learning (5,000 yen)
Recently talked about deep learning. A book by ian Dai-sensei, the developer of the particularly hot generative model "GAN". If you can read and understand "Deep learning from scratch", I would recommend this book for the next one. A book that advances while learning the knowledge of analysis and linear algebra necessary for deep learning in order.
Even if you can't read it completely, the cover is beautiful, so it may be possible to make it an interior.
Original: Deep learning
5th book
Japanese title: Statistical causal reasoning -model, reasoning, guessing- (10,000 yen)
I haven't read much yet, but the book of the author who developed causal reasoning and won the Turing Award. What is causal reasoning? When I searched to buy a book, I found this book and Iwanami Data Science.
Is it really a variable that should be included in the model when creating a statistical model? Does it really affect the forecast? A book you want to read when thinking about.
I would like to learn general points about model selection and linear regression in the third book, "Introduction to Statistical Learning with R", and understand the relationships between variables in this book.
Original: casuality models reasoning and inference
The Japanese translation is the first edition, and the original image of this is the second edition
This time, update as appropriate
There are many other free books on data science (in English). Study English before data science
(If you look for it, there are some suspicious things that the official is out ...)
I would like to summarize if you can introduce youtube version, MOOC version, foreign book limited edition, etc.