Introducing the book "Creating a profitable AI with Python" that allows you to learn machine learning in the shortest course

Introduction

He is the author of the book "** Creating a profitable AI with Python **". This article introduces the features of this book.

Amazon link (book) https://www.amazon.co.jp/dp/4296106961

Amazon Link (Kindle) https://www.amazon.co.jp/dp/B08F9P726T

This book support site (Github) https://github.com/makaishi2/profitable_ai_book_info/blob/master/README.md

First of all, please see the table of contents below.

Table of Contents

The title is often misunderstood, but as you can see from the table of contents, it is a serious book. It is not a book "Let's make a lot of money with Forex and stocks using AI", so please do not misunderstand.

Main target audience

This book is intended for two main readerships:

** Business expert ** A leader-class manager who is in charge of actual work, not the IT department. People who want to move their work to AI, but don't know where to start. In this case, after reading Chapters 1, 2 and 3 of this manual, read the relevant section of Chapter 5 by focusing on ** processing patterns ** that may be applicable to your business. Chapter 4 is optional, but I would like you to read verses 4.1 and 4.4 as much as possible. Chapter 6 is important, so please read it.

** Aspiring Data Scientist ** People who are originally good at programming and want to improve their skills as a data scientist. In this case, do not skip it and read it all from the beginning. The sample coding in the text of this manual is written on the assumption that you have some understanding of languages and libraries such as Python, Pandas, and matplotlib. If you don't have enough knowledge about this, please read "** Course 2 Introduction to Python for Machine Learning **" at the end of this book for a brief explanation of these libraries. The shortest course explains the minimum required functions that appear in the coding of this book. You may not have Python experience before the library, but don't worry. For these people, we have published [Introduction to Python](https://github.com/makaishi2/profitable_ai_book_info/blob/master/docs/C1-python grammar explanation.md # C11) on the support page (Github) of this book. did. This is also a ** introduction to Python in the shortest course **, narrowed down to the minimum necessary grammar that appears in the practice of the book.

Book features

** For business professionals ** For business professionals who have never had contact with AI, AI tends to have the image of "something unknown" or "a great thing that can do anything." In fact, what you can do with current AI technology is quite limited. By reading through this book, you can understand what AI can do. What you can do with AI, such as "** classification ", " regression ", and " clustering ", which are called " processing pattern **" in this book, and how to handle your own work You will be able to attach it. This is the first step ** to promote AI. The figure below is a ** processing pattern list ** summarized in Chapter 5 of this manual.

You don't have to understand the meaning of coding line by line, so be sure to follow the Python practice for the processing patterns you want to apply in your business in Chapter 5. By understanding the processing flow through Python code, you can ** imagine the application pattern of the machine learning model more concretely **. For example, the figure below is a graph that overlays the correct answer with the prediction result of the model that appears during the practice of the example (regression) in Section 5.2.

Chapter 6 also explains ** the pitfalls ** that tend to fall into the process of AI. ** PoC ** (abbreviation of Proof of Concept, which is the first technical verification to start AI) that can be applied to actual business for the first time by fully understanding what is written here. Will definitely be available for selection. The specific contents of Chapter 6 are as follows.

** For aspiring data scientists ** The biggest feature of this book for aspiring data scientists is that ** you can understand what to do to build a machine learning model in Python in the shortest course ** (pointing to this point on SNS Some readers have described it as "** Super Explosive Curriculum **").

To achieve this goal, the mathematical explanations of algorithms, which tend to take up many pages in traditional machine learning manuals, are ** limited to the minimum of figure-heavy images ** and mathematical formulas. Is only addition, subtraction, multiplication and division. Then, I focused on ** specifically how to create a machine learning model **. This lowers the mathematical hurdles often found in traditional machine learning and data science books. While omitting the difficult math story, I devoted section 1 (Section 4.4) to explain ** evaluation **, which is important in supervised machine learning, in considerable detail (specifically, ** Accuracy). ) **, ** Precision **, ** Recall , etc.). As a mathematical formula, it is a section that can be understood only by a fractional formula, so please make sure that you fully understand this section. I think that traditional introductory books for data scientists and programmers tended to end with a technical explanation of the model and implementation code, and tended to have less explanation from a business perspective. In Chapter 5 of this book, I always try to explain at the beginning that " This processing pattern can be used in this part of such work ". Aspiring data scientists will be able to understand the implementation code, including " Processing pattern and business association **", by reading the implementation code with this explanation at the beginning of the chapter in mind.

** Other features ** Other features of this manual are as follows.

• ** Training code is premised on Google Colab ** Google Colab is a Python (Jupyter Notebook) environment on the cloud that you can use immediately without any setup steps as long as you have a Gmail account. Since all the training code itself is published on the Internet (Github), it has been troublesome to install Python and Jupyter Notebook. Readers who have been unable to practice books due to difficulty can immediately move the practice code. For the specific procedure, please refer to the separate article on qiita. How to run the training code of the book "Making profitable AI with Python" on Google Colaboratory

• ** Adoption of the latest technology ** We use XGboost for regression in 5.2, Prophet for time series analysis in 5.3, and incorporate the latest machine learning technology. In this case as well, we do not go into advanced usage methods and keep it at a level that users with little experience can use immediately, so there is no worry that it will be difficult because it is the latest technology.

• ** Association analysis is also a case study ** Association analysis (a type of unsupervised learning), which is often used in marketing analysis, is an area where there was not much practical guide in Python because there is no library in scikit-learn, which is the de facto standard for Python machine learning. (The R language was often used). In this book, the practice in this area is also realized by using scikit-learn and another library mlextend. For the outline, I wrote another article on qiita, so please refer to it if you are interested. Association analysis with Python

For those who want to know more

On the support page of this book, ** Additional examples ** with the same writing style as Chapter 5 and ** Introduction to Python **, which explains Python grammar, are available. If you read this, you will get a better idea of the image of this book.

• [** Additional Case 1 ** Automation of Defect / Disease Judgment (Binary Classification and Recall)](https://github.com/makaishi2/profitable_ai_book_info/blob/master/docs/ Additional Case 1.md)
• [** Additional case 2 ** Discrimination by image (multi-value classification of image data)](https://github.com/makaishi2/profitable_ai_book_info/blob/master/docs/Additional case 2.md)
• [Introduction to C1.1 Python](https://github.com/makaishi2/profitable_ai_book_info/blob/master/docs/C1-python grammar explanation.md # C11)