"Turn" machine learning best Getting started learning materials summary

The best introductory learning material for machine learning summarizes the high-quality learning resources recommended for beginners in machine learning to help beginners get started quickly.

This article is really hard to write because I want it to really help beginners. With a blank piece of paper in front of me, I sat down to ask myself a question: what are some of the best libraries, tutorials, papers, and books that I would recommend for beginners in the face of a completely unfamiliar machine learning field?

The choice of resources is very tangled, I have to try to learn from a machine learning programmer and Beginner's perspective on what resources are best for them.

I have selected the best learning materials for each type of resource. If you are a real beginner and interested in learning in the field of machine learning, I hope you will find something useful in it. My advice is to pick one of the resources, a book, or a library, read from beginning to end, or complete all the tutorials. After choosing one, stick to your study, and then choose another resource to learn the same way after you have mastered it completely. Let's start now.

Library

I believe in a sentence: after learning to a certain extent, you need to start trying to do things. That's how I Learned to program, and I'm sure most of the other people have learned the same thing. Be aware of your limits and inspire your energy. If you know how to program, you can quickly dive into machine learning. Then make a plan to learn the mathematical knowledge of the technology before you implement an engineering system.

Find a library, read its documentation first, and then try to do something with the guide. Below is the best open source code for machine Learning Library. I don't think these libraries are suitable for your project, but they are ideal for learning, developing and modeling.

Start by selecting a library of the languages you are familiar with, and then try other more powerful libraries. If you are a good programmer, you should know that you can easily switch from one language to another. The logic of the program is the same, just the difference between the syntax and the API.

R Project for statistical Computing (R Engineering for statistical calculations): This is a software environment that uses the Lisp scripting language. Provides all the stats you want, including great drawings. The CRAN (third-party machine learning package) machine Learning category has the code written by experts in the field, with the latest interface methods and other features you can think of that can be found above. If you want to quickly model and develop, R engineering is a must-learn. But you don't have to start with the project from the beginning.

WEKA: A data mining platform that provides APIs, a number of command lines, and an image-based user interface for the entire data mining lifecycle. You can prepare data for visual development, create classifications, regressions, cluster models, and algorithms provided by many embedded and third-party components. If you need to work on a Hadoop platform, then Mahout is a good machine learning Java framework that is not related to Weka. But if you are a novice in big data and machine learning, stick to Weka and remember to learn only one thing at a time.

Scikit Learn: Machine learning Python library, relying on numpy and scipy libraries. If you are a python or ruby programmer, this library is more appropriate. The library has a friendly, powerful and well-documented support. If you want to try something else, then Orange will be a good choice.

Octave: If you are familiar with MATLAB or if you are a numpy programmer and are trying to find something different, consider Octave. It provides a data computing environment similar to MATLAB, and can be easily programmed to solve linear and nonlinear problems, which are the basis of most machine learning algorithms. If you have an engineer background, you can start learning from here.

BIGML: Maybe you don't want to do any programming, so you can use all the tools, like Weka. You can take a step further and use services like BIGML, BIGML provides a web-based machine learning interface, and the development and creation of models can all be done on the browser.

Choose one of the platforms for the practical practice of machine learning. Don't just look, do.

Video Courses

It is now popular to learn machine learning by watching videos. I watched a lot of machine learning videos on YouTube and videolectures.net, and the risk of watching videos is that you can easily just look at them and not practice them. I suggest that you take notes when you watch a video, even if you quickly throw it away. At the same time, it is recommended that you try to do whatever you are learning.

Frankly speaking, I have seen the video, no video is particularly suitable for beginners, I mean the real beginner. They are based on the assumption that the reader has the most basic knowledge of linear algebra and probability theory. The Stanford University Andrew Ng Tutorial is probably the best for getting started, and I recommend some disposable videos.

Stanford Machines Learning (Stanford machine Learning): Through Coursera can be obtained, Andrew ng presenter. In addition to enrolling, you can see all the lessons at any time and download all the handouts and class notes from the Stanford CS229 Course (Stanford CS229 Course). courses include homework, testing. The course focuses on linear algebra, using the octave environment.

Caltech Learning from Data: can be accessed in edx, Yaser Abu-mostafa presenter. All courses and materials are available on the Caltech website. As with the Stanford course, you can do your homework and tasks at your own pace. It covers courses similar to Stanford, and then has some depth in the details and uses more knowledge of mathematics. Homework may be too difficult for beginners.

Machine learning category on Videolectures.net (videolectures.net): Beginners can easily indulge in a huge amount of content. You can look for some interesting videos and try them out. If it's not what you can understand at this stage, let it go. If you look right, take notes. I found myself constantly searching for the labels I was interested in, and then finally choosing a completely different label. Of course, it's good to see what the experts in the field are really like.

"Getting in Shape for the Sport of Data science" –talk by Jeremy Howard: and a local R user team on machine learning Practice application dialogue, the team in the machine learning competition has achieved very good results. This video is useful because very few people talk about what it really is like to apply machine learning to a project and how to do it. I imagined that I could create a live TV show so that I could see the performance of the player in the machine learning contest directly. How much I yearn for AH.

Paper overview

If you are not used to reading research papers, you will find their language very dull. The paper is like a snippet of a textbook, but the paper describes the experiment, or other frontier research in the field. However, if you are about to start learning machine science, here are some papers that might interest you.

The discipline of machines learning (principle of machine learning): A white Paper on the definition of machine learning principles, author Tom Mitchell. There was a debate, and Mitchell eventually persuaded President CMU to set up a separate machine learning facility to ensure that machine learning would exist as a discipline for the next 100 years. (You can also refer to the short film Tom Mitchell interview).

A Few useful things to Know on machine learning (some things you have to know about machines learning): This is a good paper, because it does not adhere to specific algorithms, but rather to such important issues as feature selection overview and model simplification. It's good to go in the right direction and think it out from the start.

I've only listed two important papers, because reading a paper can really get you into trouble.

Machine Learning Beginner Books

There are many machine learning books on the market, but few are written for beginners. What is a real beginner? It may have been transferred from other fields to machine learning, or from computer science, software programming, or statistics. Even so, most books will assume that you have at least the knowledge background of linear algebra and probability theory.

However, there are some books that encourage interested programmers to start with a minimal algorithm, specify tools and libraries so programmers can run programs and get results. The most famous are programming collective Intelligence (Chinese version: "Collective Intelligence Programming"), machine learning for Hackers (Chinese version of "Computer Learning: Practical Case Analysis") and data Mining: Practical machine learning Tools and techniques (Chinese version of Data Mining: Utility learning Technology), the above three are based on Python,r and Java three languages respectively. If you do not know the place, you can see these three books.

[! [Books for machine learning beginners] (http://machinelearningmastery.com/wp-content/uploads/2013/11/photo-300x225.jpg)] (http://machinelearningmastery.com/wp-content/uploads/2013/11/photo.jpg)

Programming Collective Intelligence (Chinese version: Collective Intelligence programming): Create a sophisticated Web 2.0 application: This book is written specifically for programmers, light theory, heavy combat, there are a lot of code examples, The actual web problem encountered and the corresponding solution. Readers of this book are advised to do exercises while reading!

Machines Learning for Hackers (Chinese: Machine Learning: Use case resolution): I suggest reading this book after reading programming collective intelligence. The book also provides a large number of practical cases. But it has more data to analyze things and uses R language. I really like this book.

Machine Learning:an Algorithmic Perspective: This book is like an upgraded version of programming collective intelligence. The two books have the same goals (to help programmers get started), but the book contains math and references, as well as examples and snippets of Python writing. I suggest that the reader first look at programming collective Intelligence, if after reading the interest, then come to see this book.

Data mining:practical machine learning Tools and techniques, third Edition (Chinese version: Mining: A Practical learning technology): Actually, I learned it from this book, The first edition of 2000. I was also a Java programmer, because the Weka library provides a good development environment, I use this book with the Weka library to try, with their own algorithm to do plug-ins and a lot of machine learning applications, while extending to the Data Mining section. So I strongly recommend this book and this learning method.

Machine Learning (Chinese): This book is older, contains formulas and a lot of references. Although it is a textbook, the practicality of each algorithm is still very strong.

Many people can enumerate a lot of excellent machine learning textbooks, I can also. These books are really great, just a person who feels that it's not very good for beginners to get started.

Extended Reading

This article was carefully scrutinized to make sure that I didn't miss anything important, and I also looked at the resources listed by others. To make the content more comprehensive, here are a list of some of the other great machine learning Starter resources on the web.

A list of Data science and machine learning resources: A very detailed resource list, take some time to read the author's suggestions, and then look at the links. It's worth a look.

What is some good resources for learning on machine learning? Why?: The first answer to this question is very good, and every time I read it I take notes and label them. The most valuable part of this answer is the list of notes for the machine learning course and the related articles in the Q&a page.

Overwhelmed by machines Learning:is there an ML101 book?: A StackOverflow Post that does list a list of recommended machine learning books, the first one to reply is Jeff Moser, A lot of videos and conversations are listed.

Have you ever seen or used these resources? What do you think?

I do not know whether I am interested in learning machine learning programmer to provide a real useful resources, have any questions and suggestions, please leave a valuable message!

"Turn" machine learning best Getting started learning materials summary