search

buckyroberts / Ella

Self-improving decision organism
447 stars 85 forks source link

readme

Introduction

Ella is a self-improving decision organism.

Overview

This program is designed to iterate through a dataset and may choose to perform an action based on analysis of the data.

Diagram of Basic Neuron

Example Usage

Ella works by first analyzing the dataset provided by the user:

[
  {"date": "1/1/17", "price": 25.48, "volume": 5500},
  {"date": "1/2/17", "price": 19.64, "volume": 1600},
  {"date": "1/3/17", "price": 25.57, "volume": 4800},
  {"date": "1/4/17", "price": 32.63, "volume": 2100},
  {"date": "1/5/17", "price": 29.85, "volume": 3700}
]

She will then begin building the neurons by first generating random conditions based on the data.

volume > 3700
price < 19.64

Once conditions have been created, they are mapped to an action (creating the full neuron):

volume > 3700 - BUY
price < 19.64 - SELL

Ella will continue to create neurons until all actions have been mapped to a condition and the brain is complete. This will ensure that the brain will have the ability to perform all possible functions.

Once full brain has been constructed, Ella will iterate over the dataset row by row and invoke all action events (fire the neuron) if a condition is met. Therefore for each row analyzed, Ella may invoke multiple, a single, or no action events depending on how many conditions were met.

Example using the above neurons and dataset:

Starting value:
$1000.00

Conditions:
volume > 3700 - BUY
price < 19.64 - SELL

---------- ANALYSIS ----------

Day 1 - {'price': 25.48, 'volume': 5500}
volume 5500 > 3700 -> BUY

Day 2 - {'price': 19.64, 'volume': 1600}
> NO ACTION

Day 3 - {'price': 25.57, 'volume': 4800}
volume 4800 > 3700 -> BUY

Day 4 - {'price': 32.63, 'volume': 2100}
> NO ACTION

Day 5 - {'price': 29.85, 'volume': 3700}
> NO ACTION

---------- RESULTS ----------

Final value:
$1008.65

Developers

Feel free to contribute any by providing any bug fixes, suggestions, ideas, or anything else that may be of help.

Also, please note that this is just a small program I wrote this morning because I was bored waiting for the hockey game to start. It will most likely not be actively maintained, however I will check back frequently for pull requests and may also add some new features in my spare time.

Links