From Deep Filtering to Deep Econometrics

ArXiv ID: 2311.06256 “View on arXiv”

Authors: Unknown

Abstract

Calculating true volatility is an essential task for option pricing and risk management. However, it is made difficult by market microstructure noise. Particle filtering has been proposed to solve this problem as it favorable statistical properties, but relies on assumptions about underlying market dynamics. Machine learning methods have also been proposed but lack interpretability, and often lag in performance. In this paper we implement the SV-PF-RNN: a hybrid neural network and particle filter architecture. Our SV-PF-RNN is designed specifically with stochastic volatility estimation in mind. We then show that it can improve on the performance of a basic particle filter.

Keywords: Stochastic Volatility, Particle Filtering, Recurrent Neural Networks (RNN), Market Microstructure Noise, Hybrid Architecture, Derivatives (Options)

Complexity vs Empirical Score

• Math Complexity: 7.0/10
• Empirical Rigor: 2.0/10
• Quadrant: Lab Rats
• Why: The paper employs advanced mathematical concepts such as stochastic differential equations (SDEs), Ito’s Lemma, and particle filter theory with heavy LaTeX notation and derivations. However, it is implemented on generated data with no backtests, real-world datasets, or statistical performance metrics cited in the excerpt.

  flowchart TD
    A["Research Goal"] -->|Improve Volatility Estimation| B["Data & Baseline"]
    B --> C["Hybrid Architecture"]
    C --> D["Training Process"]
    D --> E["Evaluation"]
    E --> F["Results"]

    subgraph A ["Research Goal"]
        A1["Problem: Estimating true volatility<br>amid market microstructure noise"]
    end

    subgraph B ["Data & Baseline"]
        B1["Dataset: Financial time series<br>with microstructure noise"]
        B2["Baseline: Standard Particle Filter<br>and ML models"]
    end

    subgraph C ["Hybrid Architecture"]
        C1["SV-PF-RNN Model<br>Stochastic Volatility + Particle Filter + RNN"]
    end

    subgraph D ["Training Process"]
        D1["Sequential Training:<br>Particle Filter + RNN Integration"]
        D2["Handling: Stochastic volatility<br>dynamics & noise"]
    end

    subgraph E ["Evaluation"]
        E1["Metrics: RMSE, AIC, Prediction Accuracy"]
    end

    subgraph F ["Results"]
        F1["Key Finding: SV-PF-RNN outperforms<br>standard particle filter"]
        F2["Outcome: Interpretable & accurate<br>volatility estimation"]
    end