nlp机器学习数据增强
“Data Scientists spend 80% of their time cleaning data and the rest 20% …..” I will answer that later!
“数据科学家花费80％的时间清理数据，其余20％……..”我稍后会回答！
I got it wrong when I first started reading about Data Science claiming that I want to be a Data Scientist! But before pursuing a career in Data Science, you should know that it is not the only track that involves working with data.
当我第一次开始阅读有关数据科学的文章时就说错了，我声称自己想成为数据科学家！ 但是在从事数据科学职业之前，您应该知道，它不是涉及数据处理的唯一途径。
There are many different tracks such as Data Engineering, Data Analysis, Machine Learning, Computer Vision, Natural Language Processing, and more! I am writing this article to provide the best resources, from my point of view, for beginners to start their learning journey in the mentioned tracks.
有许多不同的领域，例如数据工程，数据分析，机器学习，计算机视觉，自然语言处理等等！ 从我的角度来看，我写这篇文章的目的是为初学者提供最好的资源，让他们开始在上述曲目中学习。
It is important to mention that there are 2 distinctions for every track; the research one and the industrial one. This article tackles the industry track. So, if you do not know the difference between the mentioned tracks, please take the time to do your own research as this is not the scope of this article.
值得一提的是，每个音轨都有2个区别； 研究一和工业一。 本文介绍了行业发展轨迹。 因此，如果您不知道上述曲目之间的区别，请花点时间进行自己的研究，因为这不在本文的讨论范围之内。
Most people, including me, start by the most famous ML course on the internet by Andrew Ng. My problem with that course is that after you spend 12 WEEKS, you will not be able to start working on a simple project. It covers the theory in the best way but when it comes to writing code and getting your hands dirty, I do not recommend it.
包括我在内的大多数人都是由Andrew Ng在互联网上最著名的ML课程开始的。 我在该课程上遇到的问题是，花了12周后，您将无法开始进行简单的项目。 它以最佳方式涵盖了理论，但是当涉及到编写代码和动手工作时，我不建议这样做。
In general, you need to learn supervised ML, unsupervised ML and an introduction to Deep Learning. You can choose one of these resources:
通常，您需要学习监督式机器学习，无监督式机器学习和深度学习入门。 您可以选择以下资源之一：
Udacity Introduction to ML course
Udacity介绍MLÇ ourse
Udacity Intro to ML Nanodegree
Udacity ML纳米级简介
Udemy ML A-Z course
Udemy ML AZÇ ourse
I recommend the Andrew Ng course and this specialization for the theory part, in case you want a more in-depth explanation.
如果您需要更深入的解释，我建议您使用吴安德(Andrew Ng)课程和该专业作为理论部分。
Also, I do recommend the ML Nanodegree but it is not a beginner-friendly. It covers more advanced topics such as the model deployment on AWS.
另外，我确实推荐ML Nanodegree，但它对初学者不友好。 它涵盖了更高级的主题，例如在AWS上的模型部署。
DL is the most trending part of ML and it is essential to learn. I recommend one of these resources:
DL是ML中最流行的部分，它是学习必不可少的。 我推荐以下资源之一：
DeepAI specialization by Andrew Ng
吴安德(Andrew Ng)的DeepAI专业化
Udacity DL Nanodegree
Udacity DL纳米度
FastAI DL foundations and for coders courses
FastAI DL基础和编码器课程
The debate on the frameworks is endless. So, whether you want to start with TensorFlow or PyTorch, you will find a lot of great resources on Coursera, Udcaity, Udemy and YouTube. Feel free to contact me if you are stuck.
关于框架的辩论是无止境的。 因此，无论您是想从TensorFlow还是PyTorch开始，您都可以在Coursera，Udcaity，Udemy和YouTube上找到很多不错的资源。 如果您遇到困难，请随时与我联系。
I cannot say that these are the best resources as the topic needs a variety of skillsets depending on the entity and the domain you are working in. But they are definitely a good place to start.
我不能说这些是最好的资源，因为该主题需要各种技能集，具体取决于您所从事的实体和领域。但是，它们绝对是一个不错的起点。
Udacity Data Engineer Nanodegree
Udacity数据工程师纳米学位
DataCamp Data Engineer with Python Track
具有Python Track的DataCamp数据工程师
DataQuest Data Engineer Track
Dataquest的数据工程师跟踪
Data Analysis skillset is required in almost all of the tracks but with different weights. There are a lot of resources for this topic and the most important thing in this track is to work on real datasets. I recommend one of these resources:
几乎所有轨道都需要数据分析技能，但是权重不同。 该主题有很多资源，而本专题中最重要的是处理实际数据集。 我推荐以下资源之一：
Dataquest Data Analyst in Python Track
Python Track中的Dataquest数据分析师
Udacity Data Analyst Nanodegree
Udacity数据分析师纳米度
DataCamp Data Analyst with Python Track
具有Python Track的DataCamp数据分析师
Data Science skillset involves the data analysis and modeling (ML) skills but in case you prefer here are some tracks:
数据科学技能集涉及数据分析和建模(ML)技能，但是如果您愿意，这里提供一些跟踪信息：
DataQuest Data Scientist in Python Track
Python Track中的DataQuest数据科学家
DataCamp Data Scientist with Python Track
具有Python Track的DataCamp数据科学家
>>The ML related courses in these tracks are not the best way to start with to learn ML, I recommend the ones I mentioned in the ML section.
>>这些课程中与ML相关的课程并不是学习ML的最佳方法，我建议在ML部分中提到的课程。
Udacity Data Scientist Nanodegree (not beginner-friendly)
Udacity数据科学家纳米学位(不适合初学者)
DataScience365 courses
DataScience365课程
Deployment is a very important part of the data stack but most probably, you will not deal with this part if you are an absolute beginner.
部署是数据堆栈中非常重要的一部分，但是如果您绝对是初学者，则很可能不会处理这部分。
The ML and DL Nanodegree covers the deployment part on AWS and there are good specializations for AWS and Google Cloud on Coursera.
ML和DL Nanodegree涵盖了AWS上的部署部分，并且Coursera上的AWS和Google Cloud具有很好的专业化。
The ability to make a computer see is a pretty tough yet exciting skill. I will try to provide some good resources to enter the world of CV.
使计算机具有可视性的能力是一项相当艰巨但令人兴奋的技能。 我将尝试提供一些好的资源来进入CV领域。
Udcaity Intro to CV course
Udcaity简历课程简介
>>This course is huge and has a heavy level of Maths and image processing techniques but you can consider it as a reference. Do not start with it :)
>>这门课程非常庞大，具有很高的数学和图像处理技术水平，但是您可以将其作为参考。 不要以它开头:)
Stanford CS231n course (playlist)
斯坦福CS231n课程(播放列表)
Udacity CV Nanodegree
Udacity CV纳米学位
DL for CV course (Arabic)
DL for CV c usse (阿拉伯文)
>>This course is conducted in Arabic by Dr.Ahmed El-Sallab and it is one of the best practical resources you will find.
>> Ahmed El-Sallab博士用阿拉伯语授课，这是您发现的最佳实用资源之一。
I do not have much experience in NLP but I will recommend some of the resources I started with. I am sure there are more and better ones.
我在NLP方面经验不足，但是我会推荐一些我开始使用的资源。 我相信会有更多更好的。
Stanford 224n course (playlist)
斯坦福224NÇ ourse (播放列表)
Udacity NLP Nanodegree
Udacity NLP纳米学位
DL for NLP course
DL为NLPÇ ourse
Most of the resources I mentioned are not free but you can get most of them for free.
我提到的大多数资源不是免费的，但是您可以免费获得其中的大多数资源。
1-Coursera:
1-Coursera：
You can simply apply for financial aid for every course. It takes about 2 weeks to get a response. Also, there is Coursera for campus if your university supports this initiative.
您可以简单地为每门课程申请经济援助。 大约需要2周的时间才能得到答复。 另外，如果您的大学支持此计划，则可以在Coursera校园学习。
2-DataCamp:
2-DataCamp：
You can get a free-two month premium subscription by creating a Microsoft account and activate the visual basic benefit. Also, if you are a university student you can get a free-three month premium subscription by applying to the GitHub student pack.
您可以通过创建Microsoft帐户并激活基本视觉效果来获得为期两个月的免费高级订阅。 另外，如果您是一名大学生，则可以通过申请GitHub学生包获得三个月的免费高级订阅。
3-DataQuest:
3- DataQuest：
There different scholarships that you can check here or by searching Google.
您可以在此处或通过搜索Google查看不同的奖学金。
4-Udacity:
4-Udacity：
You can check all the scholarships through this link and you can subscribe to the mailing list for future openings.
您可以通过此链接查看所有奖学金，还可以订阅邮件列表以备将来使用。
If you are based in Egypt, there are different initiatives such as:
如果您位于埃及，则可以采取以下不同的举措：
NTL
NTL
AAL
AAL
FWD
前轮驱动
There is no single best resource to learn one of these tracks but it is about the resource that will minimize the time you will spend searching Google for the stuff you do not understand or want to learn more about. You will have to do that!
没有最好的资源来学习其中的一种方法，但这是一种资源，它可以最大程度地减少您在Google上搜索不了解或想了解的东西的时间。 您将必须这样做！
All these resources are based on my experience and it is totally fine if you do not like any of them. For example, you might prefer the text-based resources or want to start learning from references. Please feel free to pick the approach that suits you the best.
所有这些资源都是基于我的经验，如果您不喜欢其中的任何一个，那将是完全可以的。 例如，您可能更喜欢基于文本的资源，或者想要开始从参考中学习。 请随时选择最适合您的方法。
“Data Scientists spend 80% of their time cleaning data and the rest 20% complaining about cleaning data.”
“数据科学家将80％的时间用于清理数据，其余20％则抱怨清理数据。”
Would love to connect with you on LinkedIn. Please feel free to reach out.
希望在LinkedIn上与您联系。 请随时与我们联系。
翻译自: https://medium.com/@MustafaAwny/data-stack-and-machine-learning-computer-vision-and-nlp-best-resources-for-beginners-4d3c5af901a8
nlp机器学习数据增强
