Meta Labeling

Deploy a full stack application that automatically downloads live Bitcoin prices from an exchange such as CoinmarketCap. Use Advanced ML forecasting methods to build a trading strategy based on ensemble modeling. Backtest strategy.

Null Hypothesis

There exists alpha in developing and deploying a custom trading bot using an ensemble model and experimenting with proprietary datasets.

This collaborative project aims to examine this hypothesis. We will build a full-stack application following software engineering and MLOPS best practices with the goal of disproving our null hypothesis.

Team:

Leonce
Indrit
Dipta
You?

Disclaimer: This is for Educational Purposes Only.

Models taken into consideration:

Moving Average Crossover Strategy: When the short-term moving average crosses above the long-term moving average, that indicates a positive trend and hence a buy. If it crosses from the top down, it is a short signal.
Minervini Trend Filter
Hedgefundie's Ultimate Adventure
Antonacci Dual Momentum
Radge ADX Day Trade
Radge BBO
Radge Weekend Trend Trader
Bensdorp books
Connors' 7 day high/low

Figure 1. Meta Labeling Architecture.

How to create architecture diagram using Drawio in VSCode

here

How to access BTC data

We use OpenBB Terminal, an incredible open-source library that provides APIs for accessing prices, news, data, and models. A comparable commercial tool is the Bloomberg Terminal.

Follow Installation instructions from OpenBB's website

How to use the code

# Step 1: Clone the repo
https://github.com/LNshuti/meta-labelling-architecture.git

# Step 2: Create an isolated environment to manage dependencies
conda env create --file=environment.yaml

# Step 3: install required Python packages

pip install -r requirements.txt

## Install the OpenBB python package from PIP
pip install openbb

# Step 4: Open openbb using the terminal command
$ openbb

# Get Historical Bitcoin Prices
$ historical --ticker BTC

Commonly used Ticker mapping

| Ticker | Name         |
|--------|--------------|
| BTC    | Bitcoin      |
| ETH    | Ether        |

References

Marcos Lopez de Prado. Advances in Financial Machine Learning. Lopez de Prado, M. (2018). Advances in Financial Machine Learning. United Kingdom: Wiley.
Chip Huyen. Designing Machine Learning Systems. An Iteative Process for Production Ready Application.

LNshuti / meta-labelling-architecture