Paper: arXiv 2405.12988

Authors: Ayush Singh, Anshu K. Jha, Amit N. Kumar

Abstract

In this paper, our focus lies on the Merton’s jump diffusion model, employing jump processes characterized by the compound Poisson process. Our primary objective is to forecast the drift and volatility of the model using a variety of methodologies. We adopt an approach that involves implementing different drift, volatility, and jump terms within the model through various machine learning techniques, traditional methods, and statistical methods on price-volume data. Additionally, we introduce a path-dependent Monte Carlo simulation to model cryptocurrency prices, taking into account the volatility and unexpected jumps in prices.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 6.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced stochastic calculus (Merton jump diffusion SDEs, compound Poisson processes) and path-dependent Monte Carlo simulations, indicating high mathematical density. It is grounded in real cryptocurrency data (BTC/ETH on Binance) and uses statistical metrics (MAE, RMSE, F1-score) for backtesting, but lacks explicit code or complete implementation details.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Predict Cryptocurrency Prices"] --> B{"Data Source"}
  B --> C["Historical Price-Volume Data"]
  
  C --> D["Modeling Approach:<br>Merton's Jump Diffusion Model"]
  
  D --> E["Parameter Estimation<br>via ML, Statistical & Traditional Methods"]
  
  E --> F["Path-Dependent<br>Monte Carlo Simulation"]
  
  F --> G["Key Findings"]
  
  G --> H["Improved Price<br>Forecasting Accuracy"]
  G --> I["Effective Modeling of<br>Volatility & Jumps"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a path‑dependent Monte Carlo simulation based on Merton's jump diffusion model to forecast cryptocurrency price distributions.

Source sentence
Finally, we introduce a path-dependent Monte Carlo simulation in Python aimed at simulating the future distribution of spot cryptocurrency prices.

Assumes 6

  • Jump arrivals follow a compound Poisson process with intensity equal to twice the drift or forecasted drift term. model
    Source sentence
    For the jump processes, we assume the jumps arrival distributed by compound Poisson process with the intensity of the jump being equal to twice the drift or the forecasted drift term.
  • The jump diffusion model consists of three components: drift, volatility, and jump process. model
    Source sentence
    The MJD model comprises three primary components: drift, volatility, and the jump process.
  • Data are sampled at an hourly frequency (1h granularity). data
    Source sentence
    Specifically, we used the Klines spot data for the BTCUSDT pair with the granularity of ‘1h’ from January 2020 to January 2023.
  • Monte Carlo simulations are run 10,000 times to achieve a stable result. statistical
    Source sentence
    We simulate each hour’s price forecast in Monte Carlo simulations 10000 times (Nsim = 10000) to achieve a stable result and take their averages as our final prediction.
Show 2 more
  • A rolling window of 60 hours is used for computing drift and volatility terms. model
    Source sentence
    For forecasting volatility, we again use similar libraries as in drift, along with the computation of expected volatility forecast with the GJR-GARCH model. We compute a total of 6 drift terms and 5 volatility terms that is used in our Monte Carlo simulations...
  • The jump intensity is derived from the forecasted drift term, implying a relationship between drift magnitude and jump frequency. model
    Source sentence
    For the jump processes, we assume the jumps arrival distributed by compound Poisson process with the intensity of the jump being equal to twice the drift or the forecasted drift term.

Data

  • BTCUSDT 1h OHLCV data · Binance market-prices public
    January 2020 to January 2023 · BTC/USDT pair
    Source sentence
    We use freely available OHLCV data from the popular cryptocurrency exchange binance.com, focusing on the BTCUSDT pair. Data for other pairs such as ETHUSDT and ADAUSDT can also be obtained...

Recognised sources: Binance

Costs

Unclear The classifier was not confident.

Finds 3

  • The pair (pct_change_negated, linear regression) achieved one of the lowest MAPE (18) and RMSE, and the highest transactional accuracy among tested drift‑volatility combinations. predictive-accuracy
    Source sentence
    Interestingly, the pair (pct_change_negated, linear regression) has one of the best MAPE 18 and RMSE on a transactional basis. Furthermore, it demonstrated the highest transactional accuracy along with signal accuracy.
  • Average number of jumps observed in Bitcoin 1‑minute data was 3.5 jumps per day over 2021‑2022. statistical
    Source sentence
    we noticed an average of 3.5 jumps per day, over the period.
  • Classification accuracy varied across drift‑diffusion pairs, with the highest reported accuracy of 0.527 for pct_change_negated with forc_vol_lr. performance
    Source sentence
    The following table shows the results of various combinations of drift and diffusion terms on classification based metrics like Accuracy, Precision, Recall, and F1-Score.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 11 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (32)

  1. Greedy function approximation: A gradient boosting machine. (2001) — link
  2. A Closed-Form Solution for Options with Stochastic Volatility with Applications to Bond and Currency Options (1993) — link
  3. The Pricing of Options and Corporate Liabilities (1973) — link
  4. Jumps in Financial Markets: A New Nonparametric Test and Jump Dynamics (2007) — link
  5. Option pricing when underlying stock returns are discontinuous (1976) — link
  6. XGBoost (2016) — link
  7. Predicting the Price of Bitcoin Using Machine Learning (2018) — link
  8. Volatility estimation for Bitcoin: A comparison of GARCH models (2017) — link
and 24 more
  1. Bitcoin: Medium of exchange or speculative assets? (2017) — link
  2. Is Bitcoin a hedge, a safe haven or a diversifier for oil price movements? A comparison with gold (2018) — link
  3. A novel cryptocurrency price trend forecasting model based on LightGBM (2018) — link
  4. Lead-Lag relationship between Bitcoin and Ethereum: Evidence from hourly and daily data (2019) — link
  5. Bitcoin price prediction using machine learning: An approach to sample dimension engineering (2019) — link
  6. Deep Learning Approach to Determine the Impact of Socio Economic Factors on Bitcoin Price Prediction (2019) — link
  7. Bitcoin price prediction using Deep Learning Algorithm (2019) — link
  8. A Jump-Diffusion Model for Option Pricing (2002) — link
  9. Bitcoin Price Forecasting via Ensemble-based LSTM Deep Learning Networks (2021) — link
  10. Long and short-term impacts of regulation in the cryptocurrency market (2021) — link
  11. GJR-GARCH Volatility Modeling under NIG and ANN for Predicting Top Cryptocurrencies (2021) — link
  12. Bitcoin price prediction using ARIMA model (2020) — link
  13. Brownian Motion and Stochastic Calculus (1998) — link
  14. Analysis of Bitcoin Price Prediction Using Machine Learning (2023) — link
  15. LSTM–GARCH Hybrid Model for the Prediction of Volatility in Cryptocurrency Portfolios (2023) — link
  16. Metaverse: Cryptocurrency Price Analysis Using Monte Carlo Simulation (2023) — link
  17. Attention-based CNN–LSTM for high-frequency multiple cryptocurrency trend prediction (2023) — link
  18. Forecasting Bitcoin prices using artificial intelligence: Combination of ML, SARIMA, and Facebook Prophet models (2023) — link
  19. Forecasting cryptocurrencies volatility using statistical and machine learning methods: A comparative study (2023) — link
  20. Predicting the cryptocurrency market using social media metrics and search trends during COVID-19 (2024) — link
  21. Comparative Analysis of Machine Learning Algorithms for Daily Cryptocurrency Price Prediction (2024) — link
  22. Profitability trend prediction in crypto financial markets using Fibonacci technical indicator and hybrid CNN model (2024) — link
  23. Probabilistic deep learning and transfer learning for robust cryptocurrency price prediction (2024) — link
  24. A Predictive Model for Digital Currencies Prices using Geometric Brownian Motion Stochastic Differential Equation: A Case Study of the Bitcoin (2024) — link

0 of 32 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.