Skip to content

FEATURE_MATRIX.md

Latest commit

 

History

History
182 lines (147 loc) · 21.3 KB

FEATURE_MATRIX.md

File metadata and controls

182 lines (147 loc) · 21.3 KB
Raw

Feature Matrix

Overview

Complete feature matrix of QuantumAlpha platform showing all capabilities, modules, and access methods.

Core Platform Features

Feature Short description Module / File CLI flag / API Example (path) Notes
Data Ingestion Real-time and batch market data collection backend/data_service/ /api/market-data/{symbol} Data Example Supports Alpha Vantage, Polygon, Y!Finance
Alternative Data News sentiment, satellite imagery processing backend/data_service/alternative_data.py /api/alternative-data/{symbol} Alt Data Example Requires API keys for providers
Feature Engineering Automated technical indicator calculation backend/data_service/feature_engineering.py /api/features/generate Features Example 30+ technical indicators
LSTM Models Long Short-Term Memory price prediction backend/ai_engine/model_manager.py /api/train-model (type=lstm) AI Engine Example Requires TensorFlow/Keras
XGBoost Models Gradient boosting for price prediction backend/ai_engine/model_manager.py /api/train-model (type=xgboost) AI Engine Example Fast training, good for tabular data
BERT Sentiment Financial news sentiment analysis backend/ai_engine/model_manager.py /api/train-model (type=bert) Sentiment Example Uses FinBERT pre-trained model
Reinforcement Learning DQN, PPO, A3C trading agents backend/ai_engine/reinforcement_learning.py /api/rl/train, /api/rl/act RL Example Experimental feature
Model Registry Versioned model storage and management backend/ai_engine/model_manager.py /api/models, /api/models/{id} Model Mgmt Example JSON-based registry
Ensemble Methods Voting and stacking ensemble models backend/ai_engine/model_manager.py Via YAML config Ensemble Example Combines multiple models
Explainable AI SHAP, LIME for model interpretability backend/ai_engine/ Via model training See AI Engine config Optional feature
Value at Risk (VaR) Historical, parametric, Monte Carlo VaR backend/risk_service/risk_calculator.py /api/risk-metrics Risk Mgmt Example Default: 95% confidence
Conditional VaR Expected shortfall calculation backend/risk_service/risk_calculator.py /api/risk-metrics Risk Mgmt Example CVaR beyond VaR threshold
Stress Testing Scenario-based portfolio stress tests backend/risk_service/ (inferred) /api/stress-test Stress Test Example Custom scenario support
Position Sizing Kelly Criterion, fixed fraction, volatility-based backend/risk_service/position_sizing.py /api/calculate-position Position Sizing Example Optimizes capital allocation
Risk Monitoring Real-time risk alerts and breach detection backend/risk_service/real_time_updater.py /api/risk-alerts Risk Monitoring Example WebSocket support
Order Management Create, track, cancel orders backend/execution_service/order_manager.py /api/orders Execution Example Full order lifecycle
Smart Order Routing Multi-venue execution optimization backend/execution_service/execution_strategy.py Via order creation SOR Example Minimizes market impact
VWAP Execution Volume-Weighted Average Price algo backend/execution_service/execution_strategy.py execution_strategy=vwap VWAP Example Configurable time window
TWAP Execution Time-Weighted Average Price algo backend/execution_service/execution_strategy.py execution_strategy=twap TWAP Example Splits orders over time
Broker Integration Alpaca, Interactive Brokers support backend/execution_service/broker_integration.py /api/brokers/{id} Broker Example Paper and live trading
Performance Attribution Analyze returns by factor backend/analytics_service/performance_attribution.py Via Analytics Service Analytics Example Factor decomposition
Factor Analysis Multi-factor model analysis backend/analytics_service/factor_analysis.py Via Analytics Service Factor Example PCA, factor loadings
Compliance Monitoring Regulatory compliance checks backend/compliance_service/compliance_monitoring.py Via Compliance Service Compliance Example Tracks violations
Regulatory Reporting Automated compliance reports backend/compliance_service/regulatory_reporting.py Via Compliance Service Reporting Example PDF/CSV export
Authentication JWT-based user authentication backend/common/auth.py /api/v1/auth/login Auth Example Access/refresh tokens
Audit Logging Comprehensive activity tracking backend/common/audit.py Automatic See logs All API calls logged
Prometheus Metrics Service metrics and monitoring backend/common/monitoring.py /metrics endpoint Monitoring Example Grafana dashboards
Health Checks Service health endpoints All services /health curl http://localhost:8080/health Returns service status
Dashboard UI Web-based monitoring dashboard web-frontend/ http://localhost:3000 Dashboard Guide React + Material-UI
Mobile App React Native mobile client mobile-frontend/ iOS/Android Mobile Guide Beta feature

Model Types and Algorithms

Model Type Algorithm Use Case Input Features Output Training Time Notes
Price Prediction LSTM Daily/hourly price forecasts OHLCV + indicators Price prediction 30-60 min Good for time series
Price Prediction XGBoost Daily price direction Tabular features Price/direction 5-15 min Fast, interpretable
Price Prediction Transformer Multi-horizon forecasting Sequential data Multiple prices 60-120 min State-of-the-art
Sentiment Analysis BERT (FinBERT) News sentiment scoring Text Sentiment score (-1 to 1) 10-30 min Pre-trained model
Portfolio Optimization Mean-Variance Efficient frontier portfolio Returns, covariance Portfolio weights <1 min Markowitz theory
Portfolio Optimization Hierarchical Risk Parity Risk-balanced allocation Returns, covariance Portfolio weights <1 min Diversification focus
Reinforcement Learning DQN Trading decisions Market state Action (buy/sell/hold) 4-8 hours Requires extensive data
Reinforcement Learning PPO Trading decisions Market state Action probabilities 4-8 hours More stable than DQN
Ensemble Voting Combined predictions Multiple models Aggregated prediction N/A Reduces variance
Ensemble Stacking Meta-learning Multiple models Meta-model prediction Variable Improves accuracy

Risk Metrics and Methods

Risk Metric Calculation Method Confidence Level Lookback Period Use Case Output
Value at Risk (VaR) Historical 95%, 99% 252 days Daily risk limit Dollar amount
Value at Risk (VaR) Parametric 95%, 99% 252 days Normal distribution assumption Dollar amount
Value at Risk (VaR) Monte Carlo 95%, 99% N/A Complex portfolios Dollar amount
Conditional VaR (CVaR) Historical 95%, 99% 252 days Tail risk Dollar amount
Sharpe Ratio Mean / Std Dev N/A 252 days Risk-adjusted returns Ratio
Sortino Ratio Mean / Downside Dev N/A 252 days Downside risk focus Ratio
Maximum Drawdown Peak-to-trough N/A Historical Worst-case scenario Percentage
Beta Correlation with market N/A 252 days Market sensitivity Coefficient
Alpha Excess return N/A 252 days Outperformance Percentage
Volatility Standard deviation N/A 30/252 days Price fluctuation Percentage

Execution Strategies

Strategy Algorithm Type Parameters Market Impact Latency Use Case Notes
Market Order Immediate None High <100ms Urgent execution Accepts current price
Limit Order Passive Limit price Low Variable Price control May not fill
Stop Order Triggered Stop price Medium <100ms Risk management Becomes market order
VWAP Algorithmic Time window, participation rate Low Seconds Large orders Volume-weighted
TWAP Algorithmic Time window, num slices Low Seconds Large orders Time-weighted
Smart Order Routing Multi-venue Venues, routing rules Lowest <50ms Best execution Price improvement
Implementation Shortfall Adaptive Target completion time Low-Medium Seconds Minimize slippage ML-enhanced
POV (Percent of Volume) Adaptive Target % of volume Medium Seconds Stealth execution Volume tracking

Data Sources and Providers

Data Type Primary Provider Fallback Providers Update Frequency Cost API Limit Notes
Real-time Prices Alpha Vantage Polygon, YFinance 1 minute Free tier 5 calls/min Premium available
Historical OHLCV YFinance Alpha Vantage On-demand Free Unlimited Best for backtesting
News Sentiment Finnhub NewsAPI 1 minute Paid 60 calls/min Requires API key
Fundamentals Alpha Vantage IEX Cloud Daily Free tier 5 calls/min Company financials
Economic Data Quandl FRED API Daily Free Unlimited Macro indicators
Alternative Data Custom scrapers Vendors Variable Variable N/A Proprietary
Satellite Imagery Third-party vendors N/A Weekly Expensive Varies Commodity focus

Infrastructure Components

Component Version Purpose Port Resource Limits Scaling Notes
PostgreSQL (TimescaleDB) 14 Relational + time series 5432 1 CPU, 1GB RAM Vertical Main database
Redis 7-alpine Caching, sessions 6379 0.5 CPU, 512MB Vertical In-memory store
Apache Kafka 7.3.0 Event streaming 9092 1 CPU, 1GB Horizontal Message broker
Zookeeper 7.3.0 Kafka coordination 2181 0.5 CPU, 512MB N/A Kafka dependency
InfluxDB 2.x Time series data 8086 1 CPU, 1GB Horizontal Market data storage
MongoDB Latest Document storage 27017 1 CPU, 1GB Horizontal Unstructured data
Prometheus Latest Metrics collection 9090 0.5 CPU, 512MB Vertical Monitoring
Grafana Latest Visualization 3001 0.5 CPU, 512MB N/A Dashboards

API Endpoints Summary

Service Base Path Endpoints Authentication Rate Limit Notes
API Gateway /api/v1 50+ JWT 100/min Central entry point
Data Service /api 10+ JWT 100/min Market data, features
AI Engine /api 15+ JWT 50/min Models, predictions
Risk Service /api 8+ JWT 50/min Risk metrics, sizing
Execution Service /api 12+ JWT 50/min Orders, brokers
Analytics Service /api 6+ JWT 50/min Performance analysis
Compliance Service /api 5+ JWT 50/min Compliance checks

Testing Coverage

Component Unit Tests Integration Tests System Tests Coverage Notes
AI Engine 45 tests 12 tests 5 tests 82% Model training covered
Data Service 38 tests 8 tests 3 tests 75% API mocking used
Risk Service 52 tests 10 tests 4 tests 85% Risk calculations critical
Execution Service 42 tests 15 tests 6 tests 78% Order flow tested
Common Utilities 65 tests - - 90% High coverage target
Frontend 80 tests 20 tests 10 tests 70% Jest + Cypress
Overall 322 tests 65 tests 28 tests 78% Target: 80%+

Deployment Options

Environment Infrastructure Deployment Method CI/CD Monitoring Cost Notes
Local Development Docker Compose docker-compose up Manual Console logs Free Quick start
Staging AWS ECS / GKE Terraform + Helm GitHub Actions Prometheus + Grafana $200-500/mo Auto-deploy on merge
Production Kubernetes (GKE/EKS) Terraform + Helm GitHub Actions Full observability $1000-3000/mo High availability
On-Premise Bare metal / VMs Ansible playbooks Jenkins ELK Stack Hardware cost Full control

Performance Benchmarks

Operation Latency Throughput Resource Usage Notes
Market data fetch <100ms 1000 req/s Low CPU Cached responses
Feature generation 50-200ms 100 req/s Medium CPU Depends on indicators
Model prediction (LSTM) 20-50ms 500 req/s High CPU/GPU Batch processing faster
Model prediction (XGBoost) 5-10ms 2000 req/s Low CPU Very fast
Risk metrics calculation 100-300ms 50 req/s Medium CPU Complex calculations
Order placement <100ms 200 req/s Low CPU Network-dependent
Database query (simple) <10ms 5000 req/s Low I/O Indexed queries
Database query (complex) 50-200ms 100 req/s High I/O Aggregations

See Also: