VANILLA JS ARTIFICIAL NEURAL NETWORK

Ghost-Net Simulator

A purely mathematical Multi-Layer Perceptron (MLP) built from scratch without APIs. Capable of live backpropagation and gradient descent in your browser.

LIVE NEURAL TOPOLOGY

Current Epoch
0
Loss / Error Rate
--

TRAINING DATASET (MACRO OUTCOMES)

The network will learn to predict the Market Probability (0 to 1) based on two normalized inputs: [Inflation Rate, Interest Rate].

Input 1 (Inf)Input 2 (Rate)Target Outcome
0.1 (Low)0.1 (Low)0.9 (Extreme Bullish)
0.9 (High)0.1 (Low)0.5 (Neutral)
0.1 (Low)0.9 (High)0.4 (Bearish)
0.9 (High)0.9 (High)0.1 (Crash)

HEURISTIC PREDICTION ENGINE

WAITING FOR TRAINING...