LSTP is a CLI application wrapped around a perceptron neural network.
Traditional perceptrons are binary classifiers, only being able to perform operations on 0 or 1 inputs and outputs.
LSTP uses a training algorithm based on an intercept in infinite dimensions, allowing it to make predictions on non-binary data.
Data does not have to be linearly separable for LSTP to make accurate predictions, unlike any other perceptron created.
This is a general machine learning model, but works best with consistently correlated, low variability data.
The data folder in this repository contains input/output data, along with trained models.
Formatted data was obtained using the Output CSV Export Utility.