HALM: A Highly Adaptive Learning Model for Quantitive Trading

• HALM: A Highly Adaptive Learning Model for Portfolio Decision Making
  • 1. Introduction
    • 1.1. Background
    • 1.2. Overall Architecture
    • 1.3. HALM
    • 1.4. Decision-making Effect on 09/10/2021
  • 2. Tutorial
    • 2.1. Environment Preparation
      • 2.1.1 Suggestion
      • 2.1.2 Install Python
      • 2.1.2.1. Windows environment
      • 2.1.2.2. MacOS
      • 2.1.2.3. Linux
    • 2.2. Install dependency python library
    • 2.3. How to run

1. Introduction

1.1. Background

There are two main problems in the existing portfolio decision-making process as follows:

• Many portfolio models are now essentially a static model
• Some current new decision models only focus on a specific type of correlation network or use a specific model for portfolio optimization, and lack the exploration of portfolio optimization

1.2. Overall Architecture

1.3. HALM

The overall workflow of the HALM model is shown below.

1.4. Decision-making Effect on 09/10/2021

Initial Funding: $1000.00

Funding for 09/10/2021: $14485.35

Earnings Multiple: About 14.49 times

2. Tutorial

2.1. Environment Preparation

If you want to do a good job, you must first sharpen your tools. It is more important to make the development environment right than anything else.

2.1.1 Suggestion

• It is recommended to use PyCharm or VSCode, please do not use IDE
• It is recommended to use Python 3.8.X or above
• It is recommended to make the working environment coexist with multiple versions and multiple environments, use virtualenv, etc. for isolation, regardless of whether Python 3.8 is used for daily work

2.1.2 Install Python

2.1.2.1. Windows environment

A. Installation

• Download the Python 3.8.X installation package and install it, assuming the installation path is C:\Python38, do not set environment variables

pip install virtualenv

B. Create a virtual environment

• Open the command line prompt, enter the project folder, take C:\Projects\example_project as an example, if the project uses Python 3.8.X, execute the following command to create a virtual environment

cd C:\Projects\example_project
virtualenv -p C:\Python38\python.exe venv

C. Use a virtual environment

• The command line prompt needs to activate the virtual environment before it can be used, including python, pip, etc.

cd C:\Projects\example_project
.\venv\Scripts\activate.bat

The IDE configures the project's Python Interpreter to C:\Projects\example_project\venv\bin\python.exe

2.1.2.2. MacOS

A. Installation

• Install Brew, ignore if already installed

/usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"

• Install python 3.8 with Brew

brew update
brew install [email protected]

• Open the terminal and execute the command to install virtualenv

pip install virtualenv

B. Create a virtual environment

• Open the command line prompt, enter the project folder, take C:\Projects\example_project as an example, if the project uses Python 3.8.X, execute the following command to create a virtual environment

cd ~/Projects/example_project
virtualenv -p python3 venv

C. Use a virtual environment

• The command line prompt needs to activate the virtual environment before it can be used, including python, pip, etc.

cd ~/Projects/example_project
source ./venv/bin/activate

The IDE configures the project's Python Interpreter to ~/Projects/example_project/venv/bin/python

2.1.2.3. Linux

A. Install a virtual environment with the Python package management tool pip

sudo pip install virtualenv

B. After the installation is complete, you can use the virtualenv command to create a virtual environment. You only need to specify the name of a virtual environment.

virtualenv venv

C. To activate the created virtual environment, use the following command

source venv/bin/activate

D. To exit the virtual environment use the following command

deactivate

2.2. Install dependency python library

pip install -r requirements.txt

2.3. How to run

from halm import HALM

a = 1000.44  # The closing price of product A on the day
b = 564.11  # The closing price of product B on the day
h_ab = [
    {"2020-11-01": [852.59, 485.44]},
    {"2020-11-02": [891.38, 933.66]},
    {"2020-11-03": [977.14, 996.52]}
]  # Historical closing prices for products A and B
portfolio = HALM(
    price_a=a, price_b=b, historical_prices=h_ab
).halm_decision()  # Portfolio investment strategy for the next trading day