Metadata-Version: 2.1
Name: CategorizerAI
Version: 0.4.3
Summary: An AI to categorize finantial transactions
Home-page: https://github.com/kode-konveyor/categorizerai
Author: Arpad Magosanyi
Author-email: mag@kodekonveyor.com
License: UNKNOWN
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: GNU General Public License v3 or later (GPLv3+)
Classifier: Operating System :: OS Independent
Description-Content-Type: text/markdown
Requires-Dist: tensorflow
Requires-Dist: keras
Requires-Dist: pandas
Requires-Dist: psycopg2-binary
Requires-Dist: winterboot (>=0.3)

# Install

pip3 install CategorizerAI

# Usage, Example

There is an example database using a dataset with US government transactions.
The Config.py in the source root is a configuration to use it.
The updater script in the same location drives data acquisition and updating.

You can set up your own setup by studying the example. Copy over the Config.py
to somewhere in PYTHONPATH, the updater script to somewhere in your path,
and change them to your needs. 

The hardest part is tuning the AI for the particular data you have.
use FIRST_LAYER_NEURONS, SECOND_LAYER_NEURONS, BATCH_SIZE and EPOCHS in Config.py
for the tuning. Bigger is not always better!

It might happen that the neural net cannot learn the data with enough accuracy.
In this case it tells you so, and exits.


