Machine learning is a bit confusing for engineers over 4x years old, there are mathematical formulas when you open a book, there is python, what is it delicious? About five years ago, I tried to learn machine learning at Coursera, but Dr. Andrew Ng was frustrated at Week 4. What is the difference between "supervised learning" and "unsupervised learning"?
** Assumed reader ** People who have another career in the IT industry and want to find or try to acquire a machine learning career but are frustrated
** About my career ** Starting with a host development engineer, we have been working on client servers, web applications and development systems for ten years. After that, my main work is network and infrastructure.
** Relationship with machine learning ** I tried to study machine learning about 5 years ago, but I didn't understand what I was doing, so I was easily frustrated. After that, read it in magazines.
Note) Data analysis, mechanical engineering, and artificial intelligence are not strictly separated in this text.
First of all, I didn't understand what machine learning was.
I decided to divide the viewpoints and proceed so that I could understand them efficiently.
** G Test ** The G test is definitely recommended for people who want to be able to talk to users about machine learning and AI in two weeks! You can efficiently learn the history, future developments, dangers, and philosophies of artificial intelligence. The following two teaching materials are recommended.
** [Is artificial intelligence beyond humans? Beyond deep learning](https://www.amazon.co.jp/%E4%BA%BA%E5%B7%A5%E7%9F%A5% E8% 83% BD% E3% 81% AF% E4% BA% BA% E9% 96% 93% E3% 82% 92% E8% B6% 85% E3% 81% 88% E3% 82% 8B% E3% 81% 8B-% E3% 83% 87% E3% 82% A3% E3% 83% BC% E3% 83% 97% E3% 83% A9% E3% 83% BC% E3% 83% 8B% E3% 83 % B3% E3% 82% B0% E3% 81% AE% E5% 85% 88% E3% 81% AB% E3% 81% 82% E3% 82% 8B% E3% 82% 82% E3% 81% AE -% E8% A7% 92% E5% B7% 9DEPUB% E9% 81% B8% E6% 9B% B8-% E6% 9D% BE% E5% B0% BE-% E8% B1% 8A / dp / 4040800206) ** ** This book should definitely be read. Data analysis, machine learning, and artificial intelligence are intricately intertwined with various terms, and even the person using them does not understand clear definitions or differences. Chapter 1 clearly defines artificial intelligence and clarifies its differences from other terms. Furthermore, in Chapters 2 to 5, various fields in artificial intelligence are organized by tracing the history of artificial intelligence. Chapters 6 and 7 have a wealth of implications about AI when talking to users in business. It takes about two hours to read, and it is a book that you should read anyway.
** [Thorough Strategy Deep Learning G Test Generalist](https://www.amazon.co.jp/%E5%BE%B9%E5%BA%95%E6%94%BB%E7%95%A5-% E3% 83% 87% E3% 82% A3% E3% 83% BC% E3% 83% 97% E3% 83% A9% E3% 83% BC% E3% 83% 8B% E3% 83% B3% E3% 82% B0G% E6% A4% 9C% E5% AE% 9A-% E3% 82% B8% E3% 82% A7% E3% 83% 8D% E3% 83% A9% E3% 83% AA% E3% 82 % B9% E3% 83% 88-% E5% 95% 8F% E9% A1% 8C% E9% 9B% 86-% E5% BE% B9% E5% BA% 95% E6% 94% BB% E7% 95 % A5% E3% 82% B7% E3% 83% AA% E3% 83% BC% E3% 82% BA-ebook / dp / B07NDVCN99) ** There are various reference books for taking the G test, but I think that this one is enough for me. As a result of searching the Internet for only the parts that I could not understand in this book and taking the test, I was able to obtain the G test. After completing the G test, you can talk to the user using similar words for the time being. In addition, you can use terms related to data analysis, machine learning, and artificial intelligence to some extent.
One area of data analysis is machine learning, and one area of machine learning is artificial intelligence. If you don't understand data analysis as a big framework, you can only be a temporary machine learning engineer. On the other hand, if you understand the field of data analysis, you can use statistical methods, machine learning methods, and artificial intelligence as needed. There is no point in using artificial intelligence for data analysis that requires Excel!
** [Book to acquire real data analysis ability](https://www.amazon.co.jp/%E6%9C%AC%E7%89%A9%E3%81%AE%E3%83% 87% E3% 83% BC% E3% 82% BF% E5% 88% 86% E6% 9E% 90% E5% 8A% 9B% E3% 81% 8C% E8% BA% AB% E3% 81% AB% E4% BB% 98% E3% 81% 8F% E6% 9C% AC-% E6% 97% A5% E7% B5% 8CBP% E3% 83% A0% E3% 83% 83% E3% 82% AF-% E6% B2% B3% E6% 9D% 91-% E7% 9C% 9F% E4% B8% 80 / dp / 4822237729) ** The great thing about this book is that design-> pre-check-> analysis method-> analysis-> evaluation and interpretation-> expression is put together in one compact book. Clearly understand the illustrations and basic problems of how to proceed from the definition of the problematic field to data analysis. Furthermore, if you create hands-on with a template created in advance with Excel, you will definitely be a beginner in data analysis.
** [Statistical analysis for 100 million people](https://www.amazon.co.jp/1%E5%84%84%E4%BA%BA%E3%81%AE%E3%81%9F % E3% 82% 81% E3% 81% AE% E7% B5% B1% E8% A8% 88% E8% A7% A3% E6% 9E% 90-% E3% 82% A8% E3% 82% AF% E3% 82% BB% E3% 83% AB% E3% 82% 92% E6% 9C% 80% E5% BC% B7% E3% 81% AE% E6% AD% A6% E5% 99% A8% E3% 81% AB% E3% 81% 99% E3% 82% 8B-% E8% A5% BF% E5% 86% 85-% E5% 95% 93 / dp / 4822273806) ** The author, Hiromu Nishiuchi, is well known for "statistics is the strongest study." However, this book is recommended in the sense of super practice. When you start to get involved in data analysis as an IT engineer, the first challenge is how to understand the problems that are occurring in your business and bring them into data analysis. Data analysis methods and practices can be carried out without much difficulty with basic IT knowledge. However, in order to bring it from business to the data analysis field, it is necessary to have business-oriented knowledge rather than IT. This book proceeds with data analysis based on actual case studies. The software used is Excel, and the methods are mainly pivot tables and statistics. By reading this book, you can understand how to put business challenges into your application below.
** [Introduction to Practical Data Analysis Process](https://www.amazon.co.jp/%E3%83%87%E3%83%BC%E3%82%BF%E8%A7%A3%E6% 9E% 90% E3% 81% AE% E5% AE% 9F% E5% 8B% 99% E3% 83% 97% E3% 83% AD% E3% 82% BB% E3% 82% B9% E5% 85% A5% E9% 96% 80-% E3% 81% 82% E3% 82% 93% E3% 81% A1% E3% 81% B9 / dp / 4627817711) ** This book is more detailed and practical than the previous two books. But that's why reading this book first will almost certainly frustrate you. Or even if you don't get frustrated, you can't understand what you're doing and you're wasting your time reading. Each of the above two books can be completed in about a week, including hands-on. Then read this book to brush up your data analysis from hobby to business level. I always have this book by my side, and I read it to come up with ideas when I get stuck.
In order to introduce a new system in-house, it is necessary to proceed as a project. If you don't make a project, you can't afford it, you don't have the staff, and you can't communicate with related departments. Even if you understand how to proceed with a system development project, clarify what is different in a data analysis project.
** [Book to understand the project of artificial intelligence system](https://www.amazon.co.jp/%E4%BA%BA%E5%B7%A5%E7%9F%A5%E8%83%BD% E3% 82% B7% E3% 82% B9% E3% 83% 86% E3% 83% A0% E3% 81% AE% E3% 83% 97% E3% 83% AD% E3% 82% B8% E3% 82% A7% E3% 82% AF% E3% 83% 88% E3% 81% 8C% E3% 82% 8F% E3% 81% 8B% E3% 82% 8B% E6% 9C% AC-% E4% BC % 81% E7% 94% BB% E3% 83% BB% E9% 96% 8B% E7% 99% BA% E3% 81% 8B% E3% 82% 89% E9% 81% 8B% E7% 94% A8 % E3% 83% BB% E4% BF% 9D% E5% AE% 88% E3% 81% BE% E3% 81% A7-AI-TECHNOLOGY-% E6% 9C% AC% E6% A9% 8B / dp / 4798154059) ** As expected, when he is over 40 years old, he has experience in IT projects multiple times. The point here is to understand what is the difference between the IT project and the machine learning project so far, what is the same point as other system development, and what is the different point. This makes it possible to smoothly introduce machine learning / artificial intelligence projects while making use of the knowledge so far. And there are two distinct differences from previous projects. There was only one book that introduced POC and KPI in an easy-to-understand manner.
This is the core of machine learning. You can choose between Python and R as the language, but for now it is recommended to learn the basics of machine learning in Python from the materials and future potential.
** Python Data Science Handbook ** If you don't know what to start with, start here. This book is available free of charge in English. When using machine learning with Python, the packages you definitely use are numpy, pandas, and matplotlib. As an introduction, you will first learn the basics of how to use each package. Then proceed to the actual machine learning package, scikit learn. The amount is neither too much nor too little, and you can proceed without spending too much time. In the machine learning chapter, mathematical formulas are also mentioned, but you should not try to understand them forcibly and focus on how to use the package for the time being.
** [Introduction to Data Analysis with Python](https://www.amazon.co.jp/Python%E3%81%AB%E3%82%88%E3%82%8B%E3%83%87%E3% 83% BC% E3% 82% BF% E5% 88% 86% E6% 9E% 90% E5% 85% A5% E9% 96% 80-% E7% AC% AC2% E7% 89% 88-% E2% 80% 95NumPy% E3% 80% 81pandas% E3% 82% 92% E4% BD% BF% E3% 81% A3% E3% 81% 9F% E3% 83% 87% E3% 83% BC% E3% 82% BF% E5% 87% A6% E7% 90% 86-Wes-McKinney / dp / 487311845X / ref = sr_1_1? adgrpid = 58778961568 & dchild = 1 & gclid = EAIaIQobChMIjafepdH06QIVB1RgCh2FHA2kEAAYAiAAEgIRNfD_BwE & hvadid = 338517691095 & hvdev = c & hvlocphy = 1009294 & hvnetw = g & hvqmt = e & hvrand = 4247298842042338526 & hvtargid = kwd-332404598096 & hydadcr = 27269_11561182 & jp-ad-ap = 0 & keywords = python% E3% 81% AB% E3% 82% 88% E3% 82% 8B% E3% 83% 87% E3% 83% BC% E3% 82% BF% E5% 88% 86% E6% 9E% 90% E5% 85% A5% E9% 96% 80 & qid = 15791702345 & sr = 8-1 & tag = googhydr-22) ** We recommend that you read the above Python Data Science Handbook before starting this book. According to one theory, it takes 80% of the effort to prepare data and 20% of the effort to implement machine learning. This book focuses primarily on data analysis techniques, not machine learning. It deals with file input / output, data type, visualization, data conversion, etc. in more detail. Therefore, knowledge of Python is essential. It takes some time to load, including hands-on. However, if you finish this one book, you will definitely be a beginner graduate of Python, and you will be able to master numpy, pandas, and matoplotlib for data analysis.
** [Machine learning starting with Python](https://www.amazon.co.jp/Python%E3%81%A7%E3%81%AF%E3%81%98%E3%82%81%E3% 82% 8B% E6% A9% 9F% E6% A2% B0% E5% AD% A6% E7% BF% 92-% E2% 80% 95scikit-learn% E3% 81% A7% E5% AD% A6% E3 % 81% B6% E7% 89% B9% E5% BE% B4% E9% 87% 8F% E3% 82% A8% E3% 83% B3% E3% 82% B8% E3% 83% 8B% E3% 82 % A2% E3% 83% AA% E3% 83% B3% E3% 82% B0% E3% 81% A8% E6% A9% 9F% E6% A2% B0% E5% AD% A6% E7% BF% 92 % E3% 81% AE% E5% 9F% BA% E7% A4% 8E-Andreas-C-Muller / dp / 4873117984 / ref = sr_1_1? adgrpid = 52270124614 & dchild = 1 & gclid = EAIaIQobChMI2NqqidL06QIVRaqWCh0upQUOEAAYAyAAEgJTvPD_BwE & hvadid = 338518266894 & hvdev = c & hvlocphy = 1009294 & hvnetw = g & hvqmt = e & hvrand = 16546251539735673174 & hvtargid = kwd-355766215106 & hydadcr = 27267_11561158 & jp-ad-ap = 0 & keywords = python% E3% 81% A7% E5% A7% 8B% E3% 82% 81% E3% 82% 8B% E6% A9% 9F% E6% A2 % B0% E5% AD% A6% E7% BF% 92 & qid = 15191702554 & sr = 8-1 & tag = googhydr-22) ** By the way, when the above two books are finished, the preparation of data and the basic machine learning are perfect. This book is the first step towards intermediate machine learning. When you start machine learning, it takes time to cleanse the data and design the features. Kaggle's Kernel is recommended for feature design in terms of practice. However, if you understand the basics of feature design in advance in this book, you can understand why the features are designed as described in the kernel. I want to reread this award as many times as I want.
Python As mentioned earlier, Python and R are the main languages for learning machine learning. I used to use R to handle statistical data. But now I definitely recommend Python for learning machine learning.
First of all, scikit-learn provided in Python is the best package for learning machine learning. And various books and online training are also offered. In addition, Python is a powerful tool in the data cleansing step that precedes machine learning.
** Python 3 taught by an active Silicon Valley engineer at Udemy ** I won't say any hassle, it's best to start here for now. Aim when the price is low.
** [Everyone's Python](https://www.amazon.co.jp/%E3%81%BF%E3%82%93%E3%81%AA%E3%81%AEPython-%E7%AC%AC4 % E7% 89% 88-% E6% 9F% B4% E7% 94% B0-% E6% B7% B3-ebook / dp / B01NCOIC2P / ref = sr_1_1? adgrpid = 56053864951 & dchild = 1 & gclid = EAIaIQobChMIl7Xf0dT06QIV1J7Cf0dT06QIV = g & hvqmt = e & hvrand = 5750020569722182631 & hvtargid = kwd-398175049877 & hydadcr = 27263_11561108 & jp-ad-ap = 0 & keywords =% E3% 81% BF% E3% 82% 93% E3% 81% AA% E3% 81% AEpython & qid = 1591703242 googhydr-22) ** A book that teaches Python from the basics. It also explains in an easy-to-understand manner areas that are difficult to understand, such as iterators, generators, and decorators.
** [Self-study programmer from the basics of the Python language to how to work](https://www.amazon.co.jp/%E7%8B%AC%E5%AD%A6%E3%83%97%E3%83 % AD% E3% 82% B0% E3% 83% A9% E3% 83% 9E% E3% 83% BC-Python% E8% A8% 80% E8% AA% 9E% E3% 81% AE% E5% 9F % BA% E6% 9C% AC% E3% 81% 8B% E3% 82% 89% E4% BB% 95% E4% BA% 8B% E3% 81% AE% E3% 82% 84% E3% 82% 8A % E6% 96% B9% E3% 81% BE% E3% 81% A7-% E3% 82% B3% E3% 83% BC% E3% 83% AA% E3% 83% BC% E3% 83% BB% E3% 82% A2% E3% 83% AB% E3% 82% BD% E3% 83% 95-ebook / dp / B07BKVP9QY / ref = sr_1_1? adgrpid = 57356695670 & dchild = 1 & gclid = EAIaIQobChMI-rrd3dT06QIVBLeWCh3PZgFj g & hvqmt = e & hvrand = 4883408293921823160 & hvtargid = kwd-362748005277 & hydadcr = 15819_11177362 & jp-ad-ap = 0 & keywords =% E7% 8B% AC% E5% AD% A6% E3% 83% 97% E3% 83% AD% E3% 82% B0% % 83% A9% E3% 83% 9E% E3% 83% BC & qid = 1591703267 & sr = 8-1 & tag = googhydr-22) ** When you start working as a programmer, you also need knowledge other than language. Read this book to get a better understanding of the tools you need as a Python programmer. You can go to work with a new company and follow conversations like "keep this Python in the shell" or "pull the necessary source code from git".
If you just use it for honest work, you don't need much math. However, understanding to some extent will help you troubleshoot and improve your code.
** Machine learning by Andrew NG ** In terms of mathematics, the fields may shift a little. However, this course by Andrew NG provides a comprehensive explanation of the algorithm, the mathematical formulas behind it, and even the Octave code. Even if you study linear algebra and vectors alone, it is useless in practice unless you understand how they are used in machine learning algorithms and how they are coded. In that respect, this course is a well-balanced combination of academic and practical work. However, the test is quite difficult. .. ..
** [Statistics is the strongest study [Mathematics]](https://www.amazon.co.jp/%E7%B5%B1%E8%A8%88%E5%AD%A6%E3% 81% 8C% E6% 9C% 80% E5% BC% B7% E3% 81% AE% E5% AD% A6% E5% 95% 8F% E3% 81% A7% E3% 81% 82% E3% 82% 8B-% E6% 95% B0% E5% AD% A6% E7% B7% A8-% E2% 80% 95% E2% 80% 95% E3% 83% 87% E3% 83% BC% E3% 82% BF% E5% 88% 86% E6% 9E% 90% E3% 81% A8% E6% A9% 9F% E6% A2% B0% E5% AD% A6% E7% BF% 92% E3% 81% AE% E3% 81% 9F% E3% 82% 81% E3% 81% AE% E6% 96% B0% E3% 81% 97% E3% 81% 84% E6% 95% 99% E7% A7% 91% E6% 9B% B8-% E8% A5% BF% E5% 86% 85-% E5% 95% 93 / dp / 4478104514 / ref = pd_lpo_14_t_2 / 355-1421831-4833134? _Encoding = UTF8 & pd_rd_i = 4478104514 & pd_rd_r = 403b98df-5f36-4991- 9852-b5281dccbe37 & pd_rd_w = nWWg9 & pd_rd_wg = fnjUY & pf_rd_p = 4b55d259-ebf0-4306-905a-7762d1b93740 & pf_rd_r = P8J892952D4AYYSWPNTJ & psc = 1 & refRID = P8J892D Hiromu Nishiuchi is here again. Anyway, it is well known that "statistics is the strongest study", but I personally recommend this. There are mathematical formulas, but the explanations are easy to understand and are written in plain Japanese, so you can understand them even if you don't have a math background. I feel that I finally understood the meaning of the inner product in the chapters "Relationship between vector inner product and Σ" and "How to use inner product in statistics".
** [Easy Learning Mathematics for Understanding Machine Learning](https://www.amazon.co.jp/%E3%82%84%E3%81%95%E3%81%97%E3 % 81% 8F% E5% AD% A6% E3% 81% B6-% E6% A9% 9F% E6% A2% B0% E5% AD% A6% E7% BF% 92% E3% 82% 92% E7% 90% 86% E8% A7% A3% E3% 81% 99% E3% 82% 8B% E3% 81% 9F% E3% 82% 81% E3% 81% AE% E6% 95% B0% E5% AD% A6% E3% 81% AE% E3% 81% 8D% E3% 81% BB% E3% 82% 93-% E3% 82% A2% E3% 83% A4% E3% 83% 8E-% E3% 83% 9F% E3% 82% AA% E3% 81% A8% E4% B8% 80% E7% B7% 92% E3% 81% AB% E5% AD% A6% E3% 81% B6-% E6% A9% 9F % E6% A2% B0% E5% AD% A6% E7% BF% 92% E3% 81% AE% E7% 90% 86% E8% AB% 96% E3% 81% A8% E6% 95% B0% E5 % AD% A6% E3% 80% 81% E5% AE% 9F% E8% A3% 85% E3% 81% BE% E3% 81% A7 / dp / 4839963525) ** [Machine learning series learned by Yaruo] Written by the author of (http://tkengo.github.io/blog/2016/06/06/yaruo-machine-learning0/). When I see this series, my husband does it, don't do it! I realized that. After finishing this series, I started studying with this more practical book. The formulas come out perfectly, but while keeping the basics firmly, I have removed the complicated parts that are not so necessary for beginners. This book is recommended for those who are frustrated by mathematics. By the way, in terms of mathematical formulas, [Bishop Book](https://www.amazon.co.jp/%E3%83%91%E3%82%BF%E3%83%BC%E3%83%B3%E8% AA% 8D% E8% AD% 98% E3% 81% A8% E6% A9% 9F% E6% A2% B0% E5% AD% A6% E7% BF% 92-% E4% B8% 8A-CM-% E3% 83% 93% E3% 82% B7% E3% 83% A7% E3% 83% 83% E3% 83% 97 / dp / 4612061224) is prominent, but it will definitely be frustrated if a beginner touches it. I will.
Well, this is probably more than enough for a beginner. However, how to create a foothold for intermediate players is a matter of crotch. Here, I would like to introduce two books that I have read and have definitely felt the level up. There are several books called good books even at this level. However, honestly, it is often too difficult, and it is a characteristic of this level that many books are frustrated on the way.
** [Deep Learning from scratch](https://www.amazon.co.jp/%E3%82%BC%E3%83%AD%E3%81%8B%E3%82%89%E4%BD % 9C% E3% 82% 8BDeep-Learning-% E2% 80% 95Python% E3% 81% A7% E5% AD% A6% E3% 81% B6% E3% 83% 87% E3% 82% A3% E3% 83% BC% E3% 83% 97% E3% 83% A9% E3% 83% BC% E3% 83% 8B% E3% 83% B3% E3% 82% B0% E3% 81% AE% E7% 90% 86% E8% AB% 96% E3% 81% A8% E5% AE% 9F% E8% A3% 85-% E6% 96% 8E% E8% 97% A4-% E5% BA% B7% E6% AF% 85 / dp / 4873117585) **
This book is definitely the one for the beginners. To be honest, I knew the word "error backpropagation" until I read this book, but I didn't understand it at all. Even if I read a book, it doesn't come to my mind. Chapter 5 of this book is also called "Error Backpropagation" and is implemented from scratch. If you choose one in this book, it's just Chapter 5. However, in order to use the contents implemented in the learning of the neural network in Chapter 3 and the neural network in Chapter 4, I would like to read this as well.
** [Theory and Practice by Machine Learning Programming Expert Data Scientists](https://www.amazon.co.jp/Python-%E6%A9%9F%E6%A2%B0%E5%AD%A6%E7% BF% 92% E3% 83% 97% E3% 83% AD% E3% 82% B0% E3% 83% A9% E3% 83% 9F% E3% 83% B3% E3% 82% B0-% E9% 81 % 94% E4% BA% BA% E3% 83% 87% E3% 83% BC% E3% 82% BF% E3% 82% B5% E3% 82% A4% E3% 82% A8% E3% 83% B3 % E3% 83% 86% E3% 82% A3% E3% 82% B9% E3% 83% 88% E3% 81% AB% E3% 82% 88% E3% 82% 8B% E7% 90% 86% E8 % AB% 96% E3% 81% A8% E5% AE% 9F% E8% B7% B5-impress-gear / dp / 4295003379) **
This book is currently being worked on. Perhaps if you read this book and understand the code, you are considered to be an intermediate graduate in terms of implementation. Formulas will come out, but if you keep the basics in mind, you can understand them, and there is no problem if you understand "Easy learning math for understanding machine learning". The important thing is to look it up as soon as you don't understand the formula. Because, after explaining the formula, I will mercilessly drop it into the code, so I will implement the single-layer perceptron from scratch. It's simple and plain code, but it requires a lot of ability to read Python code. Using Chapter 2 as a foothold, we will proceed with other machine learning classification algorithms, data preprocessing, and cross-validation. It is a good book that you can feel the intention to explain difficult contents with plain sentences and codes as a whole.
Recommended Posts