[See Description] Building Machine Learning Framework - Python for Finance 14

TL;DR

This tutorial explores how to incorporate machine learning into a simple long short strategy in finance using Python.

Transcript

what is going on everybody welcome to part 14 of our finance with Python tutorial series in this video we're gonna be talking about is including machine learning into this kind of basic long short strategy we're we're pretty much just going to impose machine learning as another requisites to getting in or out of a company and that's basically it so... Read More

Key Insights

• 🎰 Machine learning algorithms, such as logistic regression and support vector machines, can be integrated into financial strategies to determine entry or exit points.
• 🚂 A feature window approach is utilized to train the machine learning model on slices of historical data to predict price movements.
• 🤢 Historical bar data is queried using the 'history' function, providing flexibility in selecting specific bars, frequencies, and data fields.
• 🎰 The default parameters of machine learning algorithms in scikit-learn are usually suitable for most cases, but understanding their parameters and behavior is beneficial.
• 😑 Pre-processing modules can be imported and used to manipulate data before feeding it into the machine learning model.
• 🏪 The 'collections' module offers useful containers like the 'deque' and 'counter' for storing and manipulating data.
• 🥺 In finance, incorporating machine learning can lead to more sophisticated and data-driven investment strategies.

Questions & Answers

Q: What is the main focus of this tutorial?

The main focus of this tutorial is to demonstrate how to include machine learning as a condition for buying or selling a company in a long short strategy.

Q: What are some important libraries and modules imported in this tutorial?

The tutorial imports libraries such as scikit-learn's logistic regression, support vector machines, and random forest classifier for implementing machine learning algorithms.

Q: How is historical data used in this strategy?

The strategy considers the last hundred days of historical data and divides it into ten-day windows. Each window is used to train the machine learning model and determine whether to buy or sell based on the price movement.

Q: How is historical data queried in this tutorial?

Historical data is queried using the 'history' function, which allows the selection of specific bars, frequency, and fields. In this case, the focus is on daily closing prices.

Summary & Key Takeaways

• This video tutorial focuses on incorporating machine learning as a requirement for entering or exiting a company in a basic long short strategy.

• The tutorial begins by importing necessary libraries and modules, such as logistic regression, support vector machines, and random forest classifier.

• It then covers concepts like historical bar data, feature windows, and querying historical data.