问题描述
似乎有太多与机器学习相关的子字段.是否有一本书或博客概述了这些不同的领域，每个领域做什么，也许如何入门以及需要哪些背景知识?
It seems like there are so many subfields linked to Machine Learning. Is there a book or a blog that gives an overview of those different fields and what each of them do, maybe how to get started, and what background knowledge is required?
推荐答案
这是我所听说过的关于机器学习的最好描述:
Here's the best description I've ever heard of Machine Learning:
甚至有人认为"所有内容算法是15年前的计算机科学，而今天是机器学习."
有关更多详细信息，我建议从一个有趣的介绍开始，例如Peter Norvig的根据数据进行理论分析，窥探DeepMind的工作a>，或者是最近发布的人工智能系列讲座的未来(我从上面引用).
For more details, I'd recommend starting out with a fun intro to what's possible such as Peter Norvig's Theorizing from Data talk, a peek at what DeepMind is doing, or more recently the Future of AI series of talks (that I quoted from above).
接下来，请与杰里米·霍华德(Jeremy Howard)的"为数据科学运动做准备.这是实际使用数据的非常实用的概述.
Next get your hands dirty with Jeremy Howard's "Getting In Shape For The Sport of Data Science." It's a great pragmatic overview of actually working with data.
玩了一段时间之后，请观看Ben Hamner的" Machine Learning Gremlins "(机器学习Gremlins )，它很好地说明了在进行机器学习时容易出错的地方.
Once you've played around a bit, watch Ben Hamner's "Machine Learning Gremlins" for a nice pragmatic disclaimer of what can easily go wrong when doing machine learning.
我写了一篇博客文章"计算您的技能"在花了几个月的时间试图理解 TrueSkill 之后，这是一个进行配对和排名的ML系统在Xbox Live上.这篇文章深入探讨了机器学习中需要进一步研究的一些基础统计数据.
I wrote a blog post "Computing Your Skill" after spending months trying to understand TrueSkill, the ML system that does matchmaking and ranking on Xbox Live. The post goes into some foundational statistics needed for further study in machine learning.
也许最好的学习方法就是尝试一下.一种方法是尝试一个听起来很有趣的 Kaggle 比赛.即使我在排行榜上的表现不佳，但在尝试比赛时我总是会学到东西.
Perhaps the best way to learn is to just try it. One approach is to try a Kaggle competition that sounds interesting to you. Even though I don't do great on the leaderboards there, I always learn things when I try a competition.
完成上述操作后，我将建议更正式的内容，例如Andrew Ng的在线课程.这是在大学级别，但可以接受.如果已完成上述所有步骤，当遇到一些困难时，您会更有动力不放弃.
After that you've done the above, I'd then recommend something more formal like Andrew Ng's online class. It's at the college level, but approachable. If you've done all the above steps, you'll be more motivated to not give up when you hit some harder things.
在继续操作时，您将了解诸如 R 及其许多包， SciPy ，交叉验证，贝叶斯思维，深入 学习和很多 更多.
As you continue, you'll learn about things such as R and its many packages, SciPy, Cross Validation, Bayesian thinking, Deep Learning, and much much more.
免责声明:我在Kaggle工作，上面的几个链接都提到了Kaggle，但我认为它们是一个不错的起点.
DISCLAIMER: I work at Kaggle and several of the above links mention Kaggle, but I believe they're a fantastic place to start.
这篇关于机器学习不知所措-是否有一本书ML101?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！