PCA + Hidden Markov Model pipeline classifies the macro environment into 4 regimes — updated every trading day from Fed liquidity, rates, credit spreads, and volatility.
Every trading day after market close, the pipeline ingests 10 macro indicators and runs the full PCA → HMM inference chain.
Pulls Fed balance sheet, Treasury yields, HY spreads, VIX, DXY, and S&P 500 from FRED and Yahoo Finance.
10 features → 4 orthogonal latent factors explaining >80% of variance. Eliminates noise and collinearity.
Gaussian HMM finds the 4 most probable macro states from the factor sequence. Outputs per-regime probabilities.
REST API, WebSocket stream, and daily dashboard. Includes persistence days, confidence, and expected regime duration.
The /v1/signals/latest endpoint returns everything your strategy needs in a single JSON response.
import requests # Get the latest macro regime signal resp = requests.get( "https://api.macropulse.live/v1/signals/latest", headers={"X-MacroPulse-Key": "mp_..."} ) signal = resp.json() regime = signal["regime"]["most_likely"] # "expansion" conf = signal["regime"]["confidence"] # "HIGH" persist = signal["regime"]["persistence_days"] # 14 liq_zscore = signal["net_liquidity"]["zscore"] # 1.2 # Adjust equity exposure by regime EXPOSURE = { "expansion": 1.00, "recovery": 0.75, "tightening": 0.25, "risk_off": 0.00, } exposure = EXPOSURE[regime] # → 1.00
The HMM identifies four distinct macro states. Each maps to a different risk posture.
Fed liquidity ample, credit spreads tight, volatility low. Full risk-on. Historically the strongest return environment.
Liquidity re-injecting after stress. Risk appetite healing. Reduced but positive equity allocation.
Fed hiking, liquidity contracting, credit spreads widening. Defensive positioning. Avoid duration risk.
Crisis mode. Emergency liquidity injections, spiking spreads and vol. Flat / short. Capital preservation first.
No credit card required to start. Upgrade when you're ready to go live.
14-day money-back guarantee · Cancel anytime · Billed monthly