What is quantitative stock analysis?

Quantitative stock analysis uses mathematical models, statistical methods, and machine learning algorithms to evaluate securities. StoQuant combines 93 features including technical indicators, fundamental ratios, sentiment signals, and macro data to generate alpha scores for 822 stocks.

How does the hidden gem screener work?

The hidden gem screener uses the Benjamin Graham intrinsic value formula (V = EPS x (8.5 + 2g)) to identify undervalued small-cap stocks. It filters for P/E under 15, P/B under 1.5, PEG under 1.0, debt-to-equity under 0.5, and flags stocks trading at least 30% below intrinsic value with fewer than 8 analyst coverage.

What machine learning models power the predictions?

StoQuant uses a 3-model stacking ensemble: LightGBM, CatBoost, and XGBoost with a logistic regression meta-learner. Predictions are calibrated with Venn-ABERS conformal prediction. The system uses 93 features spanning technical, fundamental, sentiment, macro, and alternative data signals.

How does portfolio optimization work?

StoQuant integrates ML price forecasts with the Black-Litterman portfolio optimization model. ML predictions serve as investor views (Q vector) with model confidence intervals mapped to uncertainty (omega matrix). This produces optimal portfolio allocations that blend market equilibrium with AI-driven insights.

Is my portfolio data secure?

Yes. StoQuant uses JWT authentication with timing-safe comparisons, rate limiting (300 req/min per IP), secure headers (HSTS, X-Frame-Options, X-Content-Type-Options), and portfolio ownership verification on every request. No sourcemaps are exposed in production.

What data sources does StoQuant use?

StoQuant aggregates data from SEC EDGAR filings, FMP financial data, social media sentiment (Reddit, StockTwits, Twitter), Google Trends, insider trading records, institutional holdings, options flow, and macro economic indicators from FRED. All data is stored as append-only time series.

Does StoQuant work with Claude Desktop or Claude Code?

Yes. StoQuant ships an official Model Context Protocol (MCP) server, stoquant-mcp, available on npm. One command — npx -y stoquant-mcp install — wires Claude Desktop, Claude Code, or any MCP-compatible client into roughly 30 read-only stock-research tools: prices, Q-Score, ML predictions, hidden-gem screener, insider trades, market regime, and macro dashboards. Power-tier subscription required.

AI stock picks with a published track record

A 93-feature LightGBM + CatBoost + XGBoost ensemble with Venn-ABERS calibration. Out-of-sample only. AUC, hit rate, and quintile spread published daily on /proof.

Real ML, not a buzzword

Most "AI stock pick" services hide behind a black box. StoQuant publishes the ensemble (LightGBM + CatBoost + XGBoost), the meta-learner (logistic regression with Venn-ABERS conformal prediction), and the validation protocol (walk-forward out-of-sample with realistic transaction costs and slippage). The model files are versioned, the AUC is tracked, and the top-5 feature importances are visible on every prediction.

How it works

Engineer 93 features per stock — Technical, fundamental, sentiment, macro, alt-data, options, forensic, and cross-sectional features — recomputed daily from append-only time series.
Stack three boosted-tree models — LightGBM, CatBoost, and XGBoost each predict 30/60/90-day excess return probability. A logistic meta-learner blends them; Venn-ABERS calibrates the output.
Walk forward, never peek — Train on a rolling window, predict on the next out-of-sample window. Real forward returns are stored append-only and matched to predictions for honest accuracy tracking.

Related on StoQuant

Understand the methodology: Q-Score Methodology (stoquant.com/learn/q-score-methodology) and Walk-Forward Validation (stoquant.com/learn/walk-forward-validation). See today's top picks at Today's Top Q-Score (stoquant.com/today/top-q-score). Compare against black-box services: StoQuant vs Tickeron (stoquant.com/compare/tickeron).

FAQ

How accurate are the AI stock picks?

Daily AUC and hit rate are published on /proof, broken down by horizon (30/60/90 days) and walk-forward window. Numbers move; the methodology does not.

What features does the model use?

Ninety-three engineered features across technical, fundamental, sentiment, macro, alt-data, options flow, forensic accounting, and cross-sectional groups. Full feature list on /methodology.

Is it really machine learning, not a rule-based screen?

Yes. A LightGBM + CatBoost + XGBoost stacking ensemble with logistic-regression meta-learner and Venn-ABERS calibration. Models are retrained on a walk-forward schedule, not hardcoded.

How is StoQuant different from other AI stock-pick services?

Open methodology, walk-forward out-of-sample validation, and a free daily Q-Score leaderboard. No paywalled commentary, no opinions, no black box.