MNN Trading System

Motor Neural Network Trading System is build based on muscular motor network. There are two major components: 1. motor neural network clusters and 2. centralized master neurons. With well designed network, it is possible to create a desired behavior heirarchy. MNN clusters are responsible for paricular behavior while CMN are responsible for modulating the system behavior.

Motor Neural Network Cluster represents a subset of motor neurons that is responsible for a limited set of actions. Adding and substracting clusters will be equivalent to adding and subtracting functions to the application

Central Master Neural Network takes information from external and internal sources to modulate the behavior of MNN-Clusters.

Environment represents the space in which the system operates. The environment is responsible for external data, and epoch management. Environment will interact with the system by inputting sensory data to the CMNN and receiving/sending data from the MNN-clusters.

• Data
  • External API
• Epoch Management
• Simulation

Optimization and testing viability of the system will be performed by simulating a fake/past environments. Because MNN system is an input-output system, MNN will behave in a deterministic manner.

Genetic Algorithm will optimize the system.

1. Set of simulation environments are chosen
2. All synaptic connections are given a randomized action potential thresholds
3. Simulate different envirnoments and measure how each system performs
4. Select set of best performing systems and repeat step 2 - 4, while decreasing the size of randomization each time
5. After a certain threshold is met, either time or performance metrics. Select the best performing system

Classes - PascalCase Functions

• External Action Function - camelCase
  • verb is written in lowercase
    • get
    • is: sub-class of get only returns a bolean value
    • set
    • add
    • remove
    • update
    • connect: specific case to create a synapse
• External Function - PascalCase

Variables

• Class Variables - PascalCase
• Other Variables - snake_case