de对冲基金_在2020年代获得顶级对冲基金工作的最佳机器学习算法

de对冲基金

Hedge Funds analyst roles represent some of the most fiercely contested roles in all of Finance (if not any industry). The work is incredibly varied and challenging, making the jobs the long term aim of many ambitious juniors from diverse backgrounds like Computer Science, finance, economics, Physics and so much more. This creates an intense competition for roles. In turn, the challenge becomes not just landing a role, but also staying relevant enough to keep the role for the long haul.

对冲基金的分析师角色代表了整个金融行业(如果不是任何行业)中最激烈竞争的角色。 这项工作的多样性和挑战性令人难以置信，这使许多背景各异的雄心勃勃的青年成为长期目标，例如计算机科学，金融，经济学，物理学等等。 这为角色竞争激烈。 反过来，挑战不仅成为一个角色，而且要保持足够的相关性以长期保留角色。

One of the areas that hedge funds are investing in, is increasing their in-house data science and machine learning talent. In a recent survey, 56% of firms said that they would be using machine learning in their trading process by 2021 (Greenwich Associates, 2019). If we dive deeper into this, we can see that adoption of “ai” in buy-side firms is lagging the rest of the industry at only 21%. However, 36% said that they would be implementing “ai” systems in the next 2 years. To do this, they’re going to have to hire lots of new people to build and manage and maintain these systems.

对冲基金投资的领域之一，就是增加内部数据科学和机器学习人才。 在最近的一项调查中，有56％的公司表示到2021年将在交易过程中使用机器学习(Greenwich Associates，2019)。 如果我们更深入地了解这一点，我们可以看到，在买方公司中采用“ ai”正仅落后于其他行业21％。 但是，有36％的人表示，他们将在未来2年内实施“ ai”系统。 为此，他们将不得不雇用很多新人来构建，管理和维护这些系统。

To the uninitiated, the answer might then be to become an expert in machine learning — full stop. For those of us who’ve got a deeper understanding of the domain, this isn’t a realistic goal. Sure, having a base understanding of different skills — deep learning, optimisation, unsupervised methods — is important. But to deliver an edge over the thousands of other candidates, you’re going to have to bring a deep understanding and expertise of a specific area. The question is, which area should you make your own?

对于初学者来说，答案可能是成为机器学习方面的专家-永不止步。 对于那些对域有更深入了解的人，这不是一个现实的目标。 当然，对不同技能(例如深度学习，优化，无监督的方法)有基本的了解很重要。 但是要想在其他数千名候选人中脱颖而出，您将必须对特定领域有深入的了解和专业知识。 问题是，您应该在哪个区域做自己的？

为什么要让NLP成为重点 (Why you should make NLP your focus)

Not wanting to beat around the bush, or drag this out, we think you should seriously consider making NLP your core expertise. There are two main reasons for this:

不想打败或拖延，我们认为您应该认真考虑使NLP成为您的核心专业知识。 这有两个主要原因：

• You will be able to produce unique alpha
  您将能够产生独特的Alpha
• You will be more adaptable and able to apply to more firms
  您将更加适应并能够适用于更多公司

Providing a unique edge

提供独特的优势

‘Machine learning deployed on stock prices will fail. Price has no memory; unstructured ML leads to overfitting, counter-intuitive forecasts, illogical clusters and spurious results. ML applied to natural language processing of CEO veracity or chat-room sentiment will prove better predictors of performance than fundamental analysts’ — Eric Sorensen CEO PanAgora Asset Management

“以股票价格部署的机器学习将失败。 价格没有记忆； 非结构化的ML导致过度拟合，违反直觉的预测，不合逻辑的簇和虚假结果。 与基本分析师相比，将ML应用于CEO真实性或聊天室情绪的自然语言处理将证明是绩效的更好预测指标-Eric Sorensen首席执行官PanAgora Asset Management

Whilst this quote is slightly sensationalist and dismissive of the successes of some quant funds, there is a lesson here too. Many thousands of incredibly talented people who’ve spent every day analysing every piece of structured data for its ability to predict stock prices, and they’ve been doing it for the last 50 years at least. This long competition means that it is not easy to win here and opportunities rarely last long. Perhaps a better way to get an edge is to do something different. RenTec, for example, found their edge by collecting high-quality big data, decades before anyone else thought to do so.

虽然这句话有些耸人听闻，对某些量化基金的成功并不屑一顾，但这里也有一个教训。 成千上万的才华横溢的人们每天都在分析每条结构化数据以预测其股价的能力，并且至少在过去的50年中一直如此。 这场漫长的竞争意味着在这里赢得胜利并不容易，机会很少持续很长时间。 也许获得优势的更好方法是做一些不同的事情。 例如，RenTec通过收集高质量的大数据而找到了自己的优势，比其他任何人都早了数十年。

NLP provides the next frontier for this to happen. Most of the information humans produce is unstructured, stored in human-readable formats such as: Annual reports, news articles, reviews, and conversations. Tapping into this at scale is something that promises to help investment managers with all different styles:

NLP为此提供了下一个领域。 人类产生的大多数信息都是非结构化的，以人类可读的格式存储，例如：年度报告，新闻文章，评论和对话。 大规模利用这一点有望帮助各种风格的投资经理：

‘Not only sophisticated hedge funds with teams of physicists and data scientists, but also for more traditional managers. For example, fundamental managers are using machine learning to analyze earnings statements and examine market moves based on past company or macroeconomic news’ — Greenwich Associates, 2018

不仅有与物理学家和数据科学家组成的精密对冲基金，而且还有更多传统管理者。 例如，基本管理人员正在使用机器学习来分析收益表，并根据过去的公司或宏观经济新闻检查市场动向” —格林威治联合会，2018年

Because NLP has the ability to process large amounts of text and surface key insights for a human to act on, it doesn’t have to be used in a pure quantitative strategy. This means that it can add value to the investment process of traditionally fundamental investors by helping them identify key information. For you, that means there are more firms that will be looking to fill these roles and therefore more opportunities for NLP skilled data scientists.

由于NLP具有处理大量文本和表面关键洞察力以供人类采取行动的能力，因此不必将其用于纯定量策略中。 这意味着它可以通过帮助传统基础投资者识别关键信息来为其增值。 对您而言，这意味着将有更多公司寻求填补这些职位，因此为NLP熟练的数据科学家提供了更多机会。

Bringing these together, there is a real possibility that NLP could unlock the next major opportunity in investing by opening up the ability to automatically analyse vast amounts of unstructured data. When this happens, it won’t just change the workflows of quant style managers, but every investment professional. This means that by learning NLP you put yourself in a position where you are a key hire and almost any investment shop.

综合考虑，NLP很有可能通过开放自动分析大量非结构化数据的能力来释放投资的下一个主要机会。 发生这种情况时，它不仅会改变量化风格经理的工作流程，而且还会改变每位投资专家的工作流程。 这意味着通过学习自然语言处理，您可以自己处在关键职位和几乎所有投资机构的位置。

如何开始学习NLP金融专业 (How you can start learning NLP for finance)

There are many free resources that can help you learn NLP from scratch. Below we’ve linked to an example of each and looked at a bit more detail at the practical approach. It’s best to find what works for you, which may be a combination of the other options.

有许多免费资源可以帮助您从头开始学习NLP。 下面我们链接了每个示例，并详细介绍了实用方法。 最好找到适合您的方法，这可能是其他选择的组合。

• Blogs (e.g. https://nlpers.blogspot.com/, https://medium.com/huggingface

  博客 (如https://nlpers.blogspot.com/ ， https://medium.com/huggingface

• Books (e.g. https://web.stanford.edu/~jurafsky/slp3/ or https://nlp.stanford.edu/fsnlp/)

  书籍 (例如https://web.stanford.edu/~jurafsky/slp3/或https://nlp.stanford.edu/fsnlp/ )

• Courses (e.g. http://cs224d.stanford.edu/syllabus.html or see Coursera.com)

  课程 (例如http://cs224d.stanford.edu/syllabus.html或参阅Coursera.com)

There are many other guides that have links to various resources, just google how to learn natural language processing or similar.

还有许多其他指南都有指向各种资源的链接，只是google如何学习自然语言处理或类似方法。

Although these resources are good to get started with, they don’t focus on using NLP for finance specifically. This might create a knowledge gap when working on real-world problems. A lot of participants in the NLP competitions on our QuantQuest platform faced the same issue. This prompted us to create some content for bridging this gap

尽管这些资源很好入门，但它们并不专门针对使用NLP进行融资。 在处理实际问题时，这可能会导致知识缺口。 在我们的QuantQuest平台上参加NLP竞赛的很多参与者都面临着同样的问题。 这促使我们创建一些内容来弥合这一差距

实用 (Getting Practical)

Within our team, we put a lot of thought into how people learn new skills. This is partly because we need to train people internally and also because learning is a big motivator for our users. We have found that the best way to teach these topics is problem-based learning. A type of learning where skills are acquired as needed in problems, as opposed to learning theory first then trying to remember it when solving problems. This has shown to be way more engaging and hugely effective in many fields.

在我们的团队中，我们对人们如何学习新技能进行了很多思考。 这部分是因为我们需要在内部进行人员培训，还因为学习是用户的主要动力。 我们发现，教授这些主题的最佳方法是基于问题的学习。 一种学习，需要根据问题中的技能获得技能，这与先学习理论然后在解决问题时试图记住它相反。 在许多领域，这已被证明更具吸引力，并且非常有效。

For an introduction to NLP in finance, we’ve created a short course that will have you solving a real analyst’s problem within a day. As you work on the problem you’ll learn about specific libraries and approaches as you need them. The aim is to get you to a point where you’ve used NLP to do something interesting so that you’re more motivated going forward.

对于NLP的财务介绍，我们开设了一个短期课程，可让您在一天之内解决真正的分析师问题。 在解决问题时，您将根据需要了解特定的库和方法。 目的是使您达到使用NLP进行有趣的操作的地步，从而使您更有动力继续前进。

Once you’ve finished you’ll be able to start automatically analysing annual reports for insights. Then pick your next project and add layers of complexity.

完成后，您将可以开始自动分析年度报告以获取见解。 然后选择您的下一个项目，并增加复杂性。

You can check it out here

你可以在这里查看

尾注： (End Note:)

Regardless of your preferred style of learning, we thoroughly hope you consider taking a closer look at practical applications of NLP to finance. At Auquan we work for top hedge funds to help them generate more profits and avoid losses. We’re using NLP more and more to solve our clients’ challenges and see a few candidates who have a strong understanding of the topic. This skills shortage can be an opportunity for you to stand out.

无论您偏爱哪种学习方式，我们都非常希望您考虑仔细研究NLP在金融领域的实际应用。 在Auquan，我们为顶级对冲基金工作，以帮助他们产生更多的利润并避免损失。 我们越来越多地使用NLP来解决客户的挑战，并看到一些对主题有深刻理解的候选人。 技能短缺可能是您脱颖而出的机会。

翻译自: https://towardsdatascience.com/the-best-machine-learning-algorithms-to-learn-for-landing-a-top-hedge-fund-job-in-the-2020s-c58b660ac27f

de对冲基金