Machine Learning Recommended Materials

Source:http://www.openlab.co/forums/thread/413856/1

1. Machine learning
Author: (US) Tom Mitchell
Publishing house: Mechanical Industry Press
Comment: Now it seems that the book may be out of date. But in that era, it was a book of epoch-being. The first chapter clearly defines what is ML: essentially, a function is approximated by a given search space and computational resources. <RET> Although today we all use statistical means to study ml, the model and optimization algorithms have a great development, the application is more multifarious, the computational power and the year is not an order of magnitude. However, it is also a fun thing to look at what ideas in this book are out of date and what ideas are still recognized.

2. The Elements of statistical learning
Author: Trevor Hastie/robert tibshirani/jerome Friedman
Publisher: Springer
Reviews: The authors are Orthodox statistical origins Daniel, said the element, in fact, not at all. 06 Just learn ml when gnawing it, did not understand, again chew again, still did not understand. After 3 years, we will continue to feel profound. Of course, there are some reasons for the Chinese translation of the tongue, so still look at the original, can be free from the author's homepage. <RET> This book is more like product review, in the introduction of a large number of algorithmic models with the advantages and disadvantages of the comparison, which of course is the appearance of the author's skill, but not suitable for beginners. The second edition adds a lot more chapters on the sparse model than before, which is highly recommended. However, the graphic model seems to have been written too briefly.

3. Pattern Recognition and machine learning
Author: Christopher M. Bishop
Publisher: Springer
Comment: Unlike the upstairs ESL, this is almost entirely a Bayesian view, the Graphic model section is naturally a highlight. However, the biggest feature is the wide coverage, written very easy to understand. Even with Bayesian this formula is relatively many, the concept of relative around the expression, Bishop still juchongruoqing, writing fluent, voluminous. So it's great for getting started.

4. Machine Learning-a Probabilistic Perspectiv
Author: Kevin P. Murphy
Publisher: the MIT Press
Reviews: Views are between the Bayesian upstairs Prml and the ESL frequentist upstairs, 1000+ page, cover surface is very wide, from the classic model algorithm to the present stochastic optimization, deep learning. Not many formulas, concise language, easy to understand. Very suitable for beginners. <RET> the downside is that it seems more like an encyclopedia. One entry after another, it is boring to read.

5. Learning from Data
Author: Yaser S. Abu-mostafa/malik Magdon-ismail/hsuan-tien Lin
Publisher: Amlbook
Comment: There is not much comment on this book, but Yaser in Caltech's class. There is a very straightforward explanation for many of the ML concepts. There was some truth to his public contempt of Adrew Ng's ML Open class on the homepage. Http://work.caltech.edu/telecourse.html

6. Foundations of machine learning
Author: Mehryar Mohri/afshin rostamizadeh/ameet Talwalkar
Publisher: the MIT Press
Reviews: Like ESL, it's also a frequentist point of view, and it's not a foundation at all. The difference is that the author is Cs origin, so write the taste more cs dot. If you have love for bound, read it.

7. Bayesian Reasoning and machine learning
Author: David Barber
Publisher: Cambridge University Press
Reviews: Thorough Bayesian. Also wood to be read.

8. Machine Learning for Hackers
Author: Drew Conway/john Myles White
Publisher: O ' Reilly Media
Reviews: Hands-on teaching how to use ML to solve applications such as spam filtering, using R.

9. Machine learning in Action
Author: Peter Harrington
Publisher: Manning Publications
Comment: or talk about how to run the ML algorithm, using Python. Can be thought of as source code + comment + experimental result map.

Data Mining
Author: Han, Jiawei; Kamber, Micheline; Pei, Jian
Comment: All right. And downstairs this book together, is the data mining textbook. This is an optional reading if you don't want to see the formula.

One. Data Mining
Author: Ian H. witten/eibe Frank/mark A. Hall
Publisher: Morgan Kaufmann
Reviews: Weka is the most famous toolkit in data mining, which is the user's manual.

Machine Learning Recommended Materials