CHAPTER 4Machine Learning Techniques
In this chapter, we will discuss several topics centered on machine learning. The rationale behind discussing it is that machine learning can be an important part of utilizing alternative data within the investment environment. One particular usage of machine learning concerns structuring the data, which is often a key step in the investment process. Machine learning can also be used to help create forecasts using regressions, such as for economic data or prices, using various factors, which can be drawn from more traditional datasets, such as market data and also alternative data. We can also use techniques from machine learning for classification, which can be useful to help us model various market regimes.
To begin with, we give a brief discussion concerning the variance-bias trade-off and the use of cross-validation. We talk about the three broad types of machine learning, namely supervised, unsupervised, and reinforcement learning.
Then we have a brief survey of some of the machine learning techniques that have applications to alternative data. Our discussion of the techniques will be succinct, and we will refer to other texts as appropriate. We begin with relatively simple cases from supervised machine learning, such as linear and logistic regression. We move on to unsupervised techniques. There is also a discussion of the various software libraries that can be used such as TensorFlow and scikit-learn.
The latter ...
Get The Book of Alternative Data now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.