stjordanis/xbrl-ai

Applying XBRL to AI

XBRL-AI

The purpuse off this project is to apply XBRL in AI and Machine learning, by usage of Python.

The project will be divided into 3 parts:

1. Generic XBRL to AI.

• xbrl_ai.py

2. XBRL in local GAAP including

• Danish GAAP og Danish IFRS (with extension) to AI
  • xbrl_ai_dk.py
• German GAAP
  • xbrl_ai_de.py

3. Sample of Machine Learning implementation based on this project

• test_xbrl_ai_dk.py
• test_xbrl_ai_de.py
• test_xbrl_ai_us.py

Why this project?

Working with machine learning basicly comes down to one thing: y = f(X), y is what we want to predict, X is the input and f is the machine learning model. Unfortunately X hardly ever fits into to f. If we want to fit e.g. an XBRL-instance into to f, we need to prepare the data. XBRL needs a good representation to fit into AI and Machine learning.

Creating a good and standardized representation of XBRL into AI and Machine learning are the main purpose of this project.

Getting started

To see an example of how one could use xbrl-ai start by creating a conda environment from the cloned yaml file:

>> conda env create -f environment.yml

now activate the environment by using

>> activate xbrl_ai

for windows, or

>> source activate xbrl_ai

for linux.

You can now run the test_xbrl_ai_dk.py.

Installation

Download the files and run the following command from the folder:

• Using pip:

>> pip install .

• Using setuptools:

>> python setup.py install