Research Methods

Every scored paper carries one or more method labels from the catalog classifier. The topic hubs answer what is this paper about; these pages answer how does it do it. Each hub lists the guides that teach the method’s evaluation pitfalls, the surveys and methodology papers to start with, the papers with public code, and the full ranked archive for the method. Papers often carry several labels (a deep-learning paper on option pricing is also stochastic calculus), so hubs overlap.

Programmatic access: filter the methods field in papers.json; see the agents page.

Econometrics & Time-Series Methods

ARIMA/GARCH families, cointegration, state-space and regime models, realized measures, and the forecasting-evaluation toolkit (Diebold-Mariano, model confidence sets) applied to returns, volatility, and order flow.

Stochastic Calculus & Continuous-Time Models

Itô calculus, SDE and jump-diffusion models, stochastic and rough volatility, PDE/BSDE pricing, optimal stopping and Hamilton-Jacobi-Bellman control: the continuous-time machinery behind derivatives pricing and dynamic portfolio theory.

Optimization Methods

Mean-variance and robust portfolio optimization, convex and integer programming, dynamic programming, stochastic control formulations, and the regularization needed to make any of them survive estimation error.

Deep Learning

Neural networks for return and volatility forecasting, LOB modeling, option pricing and deep hedging, generative market models, and representation learning on financial sequences: LSTMs, transformers, CNNs, GNNs, and autoencoders.

Machine Learning (Non-Deep)

Tree ensembles, gradient boosting, kernel and penalized regression, clustering, feature selection and the cross-validation discipline for financial prediction, factor construction, and risk classification.

Networks & Graph Methods

Financial networks, systemic-risk contagion, correlation and minimum-spanning-tree structure, graph neural networks on asset and transaction graphs, and the network analysis of DeFi and interbank markets.

NLP & Large Language Models

Text as a signal: sentiment and embeddings from news, filings and social media, LLM agents for trading and research, retrieval over financial documents, and the evaluation problems unique to language models in finance.

Simulation & Agent-Based Models

Monte Carlo methods, agent-based market simulators, synthetic order books and price paths, generative market models, and the simulation engines used to stress-test strategies and pricing models.

Econophysics & Complex Systems

Statistical-physics tools applied to markets: random matrix theory, multifractality and Hurst exponents, scaling laws, price-impact universality, entropy measures, and the stylized facts of financial time series.

Reinforcement Learning

Deep RL for trading, execution, market making and portfolio management: DQN, PPO, actor-critic and their finance-specific reward shaping, simulators, and the evaluation problems that make most published results hard to trust.

Game Theory & Mean-Field Games

Strategic interaction in markets: Nash and Stackelberg equilibria, mean-field games for crowded trading and liquidation, auction and market-design models, principal-agent contracts, and allocation rules such as Shapley values.

Bayesian Methods

Bayesian inference and decision theory in finance: hierarchical and shrinkage priors for returns and covariances, MCMC and variational estimation of volatility models, Gaussian processes, Bayesian optimization, and posterior-based uncertainty for strategy evaluation.

Causal Inference

Event studies, difference-in-differences, instrumental variables, Granger and structural causal discovery, and causal machine learning applied to market events, policy changes, and the question of whether a predictive signal is actually causal.

Quantum Computing Methods

Quantum and quantum-inspired algorithms for portfolio optimization (QUBO, annealing, QAOA), quantum Monte Carlo and amplitude estimation for pricing, and quantum machine learning on financial data.

Learning paths

📬 The Quant Space Weekly