As an IT staff member at a business company, I was quite worried because there was no compact and comprehensive books and information that could be used in the business of the business company. There are a lot of books and information in the bookstore and on the internet, but the actual source code is scarce and the information is quite fragmented, not "exhaustive information" (this is important!) And difficult to understand. There were many things. I compared it with the curriculum of a programming school, but I thought that it would be better to read through these three books and put them into practice. If you go to a large bookstore and want to thoroughly compare the books and learn them quickly (in a few days in my image), here are the following three books. I actually bought it, but in a few days I was able to read through and grow to a level where I could write basic machine learning and deep learning code. It should be a must-have for beginners to quickly become independent as a comprehensive AI full-stack engineer!
** (1) Ready to use! Can be practiced in business! How to create AI / machine learning / deep learning apps using Python https://amzn.to/2OSFeSh**
This book is definitely a push! I immediately got an image of its application in practice. Not only the explanation of Google Colaboratory but also the code that can be actually copied and played on Colaboratory is attached in almost all exercises (it is assumed that you have built a Jupyter Notebook environment at hand). NumPy, SciPy, Pandas, scikit-learn (learning and test case division), OpenCV (OCR, handwriting and image recognition, blurring, video analysis), natural language processing, application to spam judgment, TensorFlow and Keras, MNIST, CNN It comes with source code that can be executed immediately after comprehensively learning about deep learning using libraries such as. It even mentions the difference in code when using TensorFlow, Keras, CNN library, etc. and when not using it, and it seems that everyone can become an AI engineer from tomorrow with this one book. https://amzn.to/2OSFeSh
** (2) Data Scientist Training Course at the University of Tokyo ~ Data analysis by moving hands with Python ~ https://amzn.to/2R4BIY1**
This is a book of information published online by the Matsuo Laboratory of the University of Tokyo with more details. It is very useful as an entry level for Python language characteristics, statistical analysis, and machine learning. Can be executed with Google Colaboratory. Most of the content is open to the public in the laboratory, and it is not impossible to study by itself, but since there is a large amount of it, I thought that it would be confusing and difficult to review with just the files on the net. It was helpful to have a book to put it in. It's not so expensive, so it's better to have a book at hand in order to save time. Https://amzn.to/2R4BIY1
** (3) Statistical picture book https://amzn.to/2R0zys3**
It looked like the most compact and most comprehensive book on the basics of statistics. I think it is a must-have for data scientists. Descriptive statistics, probability distributions, speculative statistics, confidence intervals, hypothesis testing, analysis of variance and multiple comparisons, nonparametrics, experimental planning, regression analysis, multivariate analysis, Bayesian statistics and big data. A keyword is comprehensively summarized in a compact manner using diagrams. https://amzn.to/2R0zys3
In my case, I thought that both the fact that I had studied some statistics in economics and the background that my knowledge of programming was not zero are enough to use even on a zero basis. If you have a little background, I think these three books are very easy to review and supplement knowledge you do not know. I think that knowledge of statistics may not be essential for AI processing, but depending on the business requirements of computer processing, it may be an area for statistical analysis, so it seems better to solidify the foundation to gain trust. ..
Recommended Posts