Jann/investment-sandbox
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Closes #019 - Live Python Machine Learning Pipeline Integration

Browse Source
This commit is contained in:
Antigravity Agent
2026-06-14 13:07:55 +02:00
parent 02cedd6ee6
commit 9c5cd78801
6 changed files with 920 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
QUANT_ROADMAP.md
View File

@@ -29,6 +29,9 @@ This document serves as the permanent, centralized system architecture design an
* **Phase 5.5: Master-Pattern KaTeX Fix & Operational Blueprint Injection**
* *Features*: Completely repaired remaining KaTeX rendering corruption in `CryptoMathModal.tsx` by converting it to the golden master single-backslash pattern. Unified math variables wrapping in text blocks across all math modals. Injected ultra-detailed operational specification matrices (Welles Wilder EMAs, Event Study isolation shields, abnormal returns, CAR, Monetization Gap, and Supply-Chain Velocity Index) directly into the operational blueprint modals for the Scanner, Econometrics, and Tech Silo modules.
* *Status*: **Fully Operational (Production Lock)**.
* **Phase 6.0: Live Python Machine Learning Pipeline Integration**
* *Features*: Integrated local Miniconda3 Python environment to automatically install `scikit-learn`. Refactored `backend/core/pipeline.py` to ingest real-time market closing price candles for BTC-USD from Yahoo Finance and funding rates from Binance USDS-M Futures REST APIs. Trained the 5 ML estimators (RF, GB, LR, SVM, MLP) across T+1, T+5, and T+10 horizons using Walk-Forward validation, exporting forecasts to `public/data/ensemble_predictions.json` with `isShieldActive: false` to enable live probabilities in the frontend Walk-Forward Radar.
* *Status*: **Fully Operational (Production Lock)**.
---